ai-consultants 5.1.0 → 5.1.2

package/.env.example

@@ -116,17 +116,19 @@ ENABLE_DEEPSEEK=false
 # =============================================================================
 
 # Gemini - The Architect (CLI mode via Antigravity CLI: agy)
-GEMINI_MODEL="Gemini 3.7 Flash (High)"
+GEMINI_MODEL="Gemini 3.8 Flash (High)"
 GEMINI_TIMEOUT=180
 GEMINI_CMD=agy
 # API-mode model ID (only used when GEMINI_USE_API=true)
 GEMINI_API_MODEL=gemini-3.1-pro-preview
 # Gemini 3.1 Pro and 3.6 Flash remain valid CLI pins. API mode stays on the
-# separately verified 3.1 Pro ID; gemini-3.7-flash is an API-only opt-in until
+# separately verified 3.1 Pro ID; gemini-3.8-flash is an API-only opt-in until
 # that transport completes its own authenticated smoke.
-# Reasoning effort for the agy CLI. Passed as --effort <value>. Accepts only
-# low|medium|high; anything else is rejected by the provider and surfaces as a
-# diagnosed error (provider owns the enum — same stance as QWEN3_REASONING_EFFORT).
+# Reasoning effort for the agy CLI, passed as --effort <value>. Current agy
+# advertises low|medium|high|max globally, but named Gemini 3.8 High/Medium/Low
+# variants reject --effort: leave GEMINI_REASONING_EFFORT unset for those CLI
+# targets and choose the level with GEMINI_MODEL. Native Gemini 3.7/3.8 Flash
+# API thinking supports low|medium|high through this setting.
 # GEMINI_REASONING_EFFORT=
 
 # Codex - The Pragmatist
@@ -160,7 +162,7 @@ KIMI_CMD=kimi
 # Claude - The Synthesizer
 CLAUDE_MODEL=claude-fable-5-1 # ai-consultants:default
 # For the lower-cost standard/medium Claude tier, set CLAUDE_MODEL to
-# claude-opus-5 (or apply_model_tier standard for a single run).
+# claude-opus-5-5 (or apply_model_tier standard for a single run).
 CLAUDE_TIMEOUT=240
 CLAUDE_CMD=claude
 CLAUDE_API_MAX_TOKENS=16384
@@ -206,7 +208,7 @@ GLM_API_URL=https://api.z.ai/api/coding/paas/v4/chat/completions
 
 # Grok Build CLI + xAI API fallback - The Provocateur
 GROK_CMD=grok
-GROK_MODEL=grok-4.6
+GROK_MODEL=grok-4.7
 GROK_TIMEOUT=180
 GROK_MAX_TURNS=4
 GROK_OAUTH_MODE=shared
@@ -215,7 +217,7 @@ GROK_API_URL=https://api.x.ai/v1/chat/completions
 # GROK_REASONING_EFFORT= # max_quality sets xhigh (Grok Build maximum)
 
 # DeepSeek - The Code Specialist
-DEEPSEEK_MODEL=deepseek-v4-pro
+DEEPSEEK_MODEL=deepseek-flash
 DEEPSEEK_TIMEOUT=180
 DEEPSEEK_MAX_QUALITY_TIMEOUT=600
 DEEPSEEK_API_MAX_TOKENS=16384

package/CLAUDE.md

@@ -6,7 +6,7 @@ AI Consultants is a multi-model coverage system that queries up to 10 AI consult
 
 **Self-Exclusion**: The invoking agent is automatically excluded from both the panel and synthesis. Claude Code won't query or synthesize with Claude, Codex CLI won't query or synthesize with Codex, etc.
 
-**Version**: 5.1.0
+**Version**: 5.1.2
 
 ## Distribution
 
@@ -349,10 +349,10 @@ Three normal tiers plus a `maximum` tier are configurable via `apply_model_tier(
 
 | Tier | Description | Example Models |
 |------|-------------|----------------|
-| **premium** | Latest flagship models | claude-fable-5-1, Gemini 3.7 Flash (High), gpt-6-astra |
+| **premium** | Latest flagship models | claude-fable-5-1, Gemini 3.8 Flash (High), gpt-6-astra |
 | **maximum** | All 10 consultants; maximum targets and highest provider effort | claude-fable-5-1, K3-256k, Qwen3.8-Max, MiniMax M3; Grok xhigh, GLM/DeepSeek max |
-| **standard** | Good quality at reasonable cost | claude-opus-5, Gemini 3.7 Flash (High), gpt-5.6-terra |
-| **economy** | Optimized for speed and low cost | claude-haiku-4-5, Gemini 3.7 Flash (Low), gpt-5.6-luna |
+| **standard** | Good quality at reasonable cost | claude-opus-5-5, Gemini 3.8 Flash (High), gpt-5.6-terra |
+| **economy** | Optimized for speed and low cost | claude-haiku-4-5, Gemini 3.8 Flash (Low), gpt-5.6-luna |
 
 **Default models are now premium tier** for maximum quality.
 
@@ -365,7 +365,7 @@ apply_model_tier "standard"  # Set all consultants to standard models
 apply_model_tier "economy"   # Set all consultants to economy models
 
 # Get model for a specific consultant and tier (v2.8.1)
-get_model_for_tier "gemini" "premium"   # → Gemini 3.7 Flash (High) on agy
+get_model_for_tier "gemini" "premium"   # → Gemini 3.8 Flash (High) on agy
 get_model_for_tier "claude" "economy"   # → claude-haiku-4-5
 ```
 
@@ -389,12 +389,12 @@ All consultants now use premium models by default:
 | Consultant | Default Model |
 |------------|---------------|
 | Claude | claude-fable-5-1 |
-| Gemini | Gemini 3.7 Flash (High) (via agy CLI); API: gemini-3.1-pro-preview |
+| Gemini | Gemini 3.8 Flash (High) (via agy CLI); API: gemini-3.1-pro-preview |
 | Codex | gpt-6-astra |
 | Mistral | CLI: mistral-medium-3.5; API: mistral-large-3 |
-| DeepSeek | deepseek-v4-pro |
+| DeepSeek | deepseek-flash (DeepSeek-V4.1-Flash) |
 | GLM | glm-5.3-flash |
-| Grok | grok-4.6 |
+| Grok | grok-4.7 |
 | Qwen3 | qwen3.7-max |
 | Kimi | kimi-code/k3 |
 | MiniMax | MiniMax-M2.7 |
@@ -402,7 +402,7 @@ All consultants now use premium models by default:
 Override with environment variables: `CLAUDE_MODEL`, `GEMINI_MODEL`, `CODEX_MODEL`, `KIMI_MODEL`, etc.
 
 Grok is CLI-first through the official Grok Build `grok` command, with
-`grok-4.6` passed explicitly in headless mode. The prompt is delivered with
+`grok-4.7` passed explicitly in headless mode. The prompt is delivered with
 `--prompt-file`; HOME and CWD are isolated, built-in/MCP tools are denied, and
 the strict sandbox prevents user or project extensions from joining the
 consultation. The xAI Chat Completions path is retained only when the CLI is
@@ -588,7 +588,7 @@ for f in scripts/*.sh scripts/lib/*.sh; do bash -n "$f" && echo "OK: $f"; done
 | `MAX_SESSION_COST` | 1.00 | Max budget ($) |
 | `CLAUDE_MODEL` | claude-fable-5-1 | Claude model (use claude-opus-5 for lower-cost standard) |
 | `CLAUDE_API_MAX_TOKENS` | 16384 | Claude API thinking + visible-output budget |
-| `GEMINI_MODEL` | Gemini 3.7 Flash (High) | Gemini agy CLI model |
+| `GEMINI_MODEL` | Gemini 3.8 Flash (High) | Gemini agy CLI model |
 | `GEMINI_API_MODEL` | gemini-3.1-pro-preview | Gemini API-mode model ID (v2.15) |
 | `CODEX_MODEL` | gpt-6-astra | Codex model (v2.5) |
 | `MISTRAL_MODEL` | mistral-large-3 | Mistral model (v2.5) |
@@ -760,6 +760,21 @@ curl -fsSL https://raw.githubusercontent.com/matteoscurati/ai-consultants/main/s
 
 ## Changelog
 
+### v5.1.2 (2026-09-28)
+
+- **Model selection remains transport-specific.** Grok 4.7 replaces the premium/maximum default, Opus 5.5 replaces standard Claude, and Gemini 3.8 High/Low replace the respective CLI targets. Fable and API Pro 3.1 remain separate defaults. Native Gemini 3.8 API thinking accepts low/medium/high and rejects incompatible explicit values before dispatch.
+- **Configuration preservation.** Exact historical defaults migrate while explicit choices survive. Quoted values plus inline default/pin markers now round-trip correctly; the previous parser retained quotes after removing the comment and missed the migration.
+- **Accounting.** Google GenerateContent output is candidates plus thoughts; prompt count already includes cached input. New rates remain estimates when no provider cost is available. Grok 4.7 rates double above 200K input; Gemini introductory rates expire on 2026-12-31 and require a catalog refresh.
+- **Validation scope.** Exact implementation-head CI passed 34/34 suites; Grok 4.7, Opus 5.5 and Gemini 3.8 High/Low CLI smokes passed on that code. Gemini named variants select effort through the model name and reject an additional --effort; initial rejected attempts and separately authorized corrected successes are recorded. Frozen P1.6 inputs and saved Grok work remain untouched. No model-quality gain is inferred from smoke checks.
+
+
+### v5.1.1 (2026-09-10)
+
+- **DeepSeek V4.1 Flash policy.** `scripts/config.sh` selects `deepseek-flash` for the default and every tier. `scripts/configure.sh` migrates only the two exact historical unpinned defaults; durable pins and explicit overrides remain intact. No credentials are imported or personal configuration rewritten by the upgrade.
+- **Conservative cost accounting.** `docs/cost_rates.json` and the catalog-independent fallbacks use $0.30/M input and $1.20/M output for Flash, including retired Flash aliases now served by the new model. `format_cost_caveats` discloses the peak/cache-miss assumption and excluded discounts; estimates are not provider invoices. Other legacy rates are retained as historical estimates.
+- **Verification and scope.** PR #27 passed all 32 CI suites, plus local API 101, configure 138 and core 451 checks. One authorized API request on `96816b5` returned HTTP 200 and a correct structured expiry-boundary answer, with `deepseek-flash` provider-reported, max effort, 119 input / 602 output measured tokens. The shared transport is unchanged; provider identity is not rewritten into a version label. No frozen P1.6 input or live benchmark claim changes.
+
+
 ### v5.1.0 (2026-09-09)
 
 - **Fable and Astra model policy.** `scripts/config.sh` and `scripts/configure.sh` align default/premium/maximum with Fable 5.1 and Astra, retain Opus 5/Terra for standard and Haiku/Luna for economy, and migrate exact historical unpinned Codex defaults while recording their origin. `resolve_codex_effort` applies implicit high per request so a later tier cannot inherit an automatically exported value.

package/README.md

@@ -1,8 +1,8 @@
-# AI Consultants v5.1.0
+# AI Consultants v5.1.2
 
 > **Coverage, not a single guess.** A panel of up to 10 frontier models from different vendors fans out on your question in parallel and hands you the *union* of what they collectively see — the risks, edge cases, and approaches a single model misses.
 
-[![Version](https://img.shields.io/badge/version-5.1.0-blue.svg)](https://github.com/matteoscurati/ai-consultants)
+[![Version](https://img.shields.io/badge/version-5.1.2-blue.svg)](https://github.com/matteoscurati/ai-consultants)
 [![License](https://img.shields.io/badge/license-MIT-green.svg)](LICENSE)
 [![Claude Code Skill](https://img.shields.io/badge/Claude%20Code-Skill-orange.svg)](https://docs.anthropic.com/en/docs/claude-code/skills)
 [![GitHub stars](https://img.shields.io/github/stars/matteoscurati/ai-consultants?style=social)](https://github.com/matteoscurati/ai-consultants)
@@ -385,7 +385,7 @@ INVOKING_AGENT=codex ./scripts/consult_all.sh "Question"    # Codex excluded
 | **Grok** | `grok` | The Provocateur | Challenge conventions |
 | **MiniMax** | `mmx` | The Pragmatic Optimizer | Performance, efficiency, pragmatism |
 
-Grok uses Grok Build with `grok-4.6` in an isolated, tool-free sandbox. Prompts
+Grok uses Grok Build with `grok-4.7` in an isolated, tool-free sandbox. Prompts
 are passed through a private file rather than process arguments. It falls back
 to the xAI API only when the CLI is missing, cannot launch, or has no usable
 authentication and `GROK_API_KEY` is configured; post-launch request failures
@@ -412,7 +412,7 @@ the observed version is response provenance only.
 | Consultant | Default Model | Persona | Focus |
 |------------|---------------|---------|-------|
 | **GLM** | glm-5.3-flash | The Methodologist | Structured approaches |
-| **DeepSeek** | deepseek-v4-pro | The Code Specialist | Algorithms, code generation |
+| **DeepSeek** | deepseek-flash (V4.1 Flash) | The Code Specialist | Algorithms, code generation |
 
 ### Installing Consultant CLIs
 
@@ -426,7 +426,7 @@ pip install mistral-vibe               # Mistral
 # Optional CLI-based consultants
 curl -L code.kimi.com/install.sh | bash            # Kimi K3
 npm install -g @qwen-code/qwen-code@latest  # Qwen (alternative to API)
-curl -fsSL https://x.ai/cli/install.sh | bash # Grok Build (grok-4.6)
+curl -fsSL https://x.ai/cli/install.sh | bash # Grok Build (grok-4.7)
 npm install -g mmx-cli                       # MiniMax
 
 ```
@@ -483,27 +483,27 @@ and cannot support a comprehensive coverage claim.
 
 | Consultant | `max_quality` | Premium | Standard | Economy |
 |------------|---------------|---------|----------|---------|
-| Claude | claude-fable-5-1 | claude-fable-5-1 | claude-opus-5 | claude-haiku-4-5 |
-| Gemini CLI | Gemini 3.7 Flash (High) | Gemini 3.7 Flash (High) | Gemini 3.7 Flash (High) | Gemini 3.7 Flash (Low) |
+| Claude | claude-fable-5-1 | claude-fable-5-1 | claude-opus-5-5 | claude-haiku-4-5 |
+| Gemini CLI | Gemini 3.8 Flash (High) | Gemini 3.8 Flash (High) | Gemini 3.8 Flash (High) | Gemini 3.8 Flash (Low) |
 | Gemini API | gemini-3.1-pro-preview | gemini-3.1-pro-preview | gemini-3.1-pro-preview | gemini-3.1-pro-preview |
 | Codex | gpt-6-astra | gpt-6-astra | gpt-5.6-terra | gpt-5.6-luna |
 | Mistral CLI | mistral-medium-3.5 | mistral-medium-3.5 | mistral-medium-3.5 | devstral-small-2 |
 | Mistral API | mistral-large-3 | mistral-large-3 | mistral-large-3 | mistral-large-3 |
-| DeepSeek | deepseek-v4-pro | deepseek-v4-pro | deepseek-v4-flash | deepseek-v4-flash |
+| DeepSeek | deepseek-flash | deepseek-flash | deepseek-flash | deepseek-flash |
 | GLM | glm-5.3-flash | glm-5.3-flash | glm-5.3-flash | glm-4-flash |
-| Grok | grok-4.6 | grok-4.6 | grok-4.5 | grok-4.5 |
+| Grok | grok-4.7 | grok-4.7 | grok-4.5 | grok-4.5 |
 | Qwen3 | qwen3.8-max when Token Plan is configured; otherwise qwen3.7-max | qwen3.7-max | qwen3.6-35b-a3b | qwen3-32b |
 | Kimi | kimi-code/k3-256k | kimi-code/k3 | kimi-code/k3 | kimi-code/k3 |
 | MiniMax | MiniMax-M3 | MiniMax-M2.7 | MiniMax-M2.7 | MiniMax-M2.5 |
 
-Promotion is transport-specific. Gemini 3.7 Flash High completed an exact live
-smoke through the `agy` adapter and is now the CLI default; Low is the CLI
-economy target. The Google API model remains `gemini-3.1-pro-preview`, while
-`gemini-3.7-flash` stays an API-only opt-in until that separate transport is
-verified. The new Mistral API IDs (`mistral-medium-3-5`,
+Promotion is transport-specific. Gemini 3.8 Flash High is the new CLI target;
+Low serves economy. Both are available in the installed `agy models` inventory;
+High/Low CLI smokes passed on 2026-09-28 with capability-probed identity. Earlier Gemini 3.7 evidence remains historical.
+The Google API default remains `gemini-3.1-pro-preview`; `gemini-3.8-flash` is
+an explicit API opt-in whose live transport is not yet verified. The new Mistral API IDs (`mistral-medium-3-5`,
 `mistral-large-2512`, `mistral-small-2603`) remain catalogued opt-ins. Claude
 Fable 5.1 is the default/premium/maximum target; use
-`CLAUDE_MODEL=claude-opus-5` for the lower-cost standard override. Selecting a preset later intentionally reapplies that
+`CLAUDE_MODEL=claude-opus-5-5` for the lower-cost standard override. Selecting a preset later intentionally reapplies that
 preset's tier and can replace an explicit model override for the run.
 Qwen3.8-Max is likewise selected by `max_quality` only when API mode already
 points at an authenticated OpenAI-compatible Token Plan `/chat/completions`
@@ -634,7 +634,7 @@ MAX_SESSION_COST=1.00        # USD budget cap (paired with ENABLE_BUDGET_LIMIT=t
 KIMI_MODEL=kimi-code/k3      # Pin the Kimi consultant to K3
 CLAUDE_API_MAX_TOKENS=16384  # Shared thinking + visible-output budget in API mode
 MISTRAL_CLI_MODEL=mistral-medium-3.5 # Vibe alias; MISTRAL_MODEL remains API-only
-GEMINI_MODEL="Gemini 3.7 Flash (High)" # agy CLI; API uses GEMINI_API_MODEL
+GEMINI_MODEL="Gemini 3.8 Flash (High)" # agy CLI; API uses GEMINI_API_MODEL
 ```
 
 Full reference: [`references/configuration.md`](references/configuration.md). Copy-paste workflows: [`docs/RECIPES.md`](docs/RECIPES.md). For category-aware preset suggestions: `ai-consultants doctor --suggest-preset --question "..."`.

package/SKILL.md

@@ -5,10 +5,10 @@ license: MIT
 compatibility: Requires bash, jq, and at least 2 AI CLI tools (agy, codex, vibe, etc.). macOS and Linux.
 metadata:
   author: matteoscurati
-  version: 5.1.0
+  version: 5.1.2
 ---
 
-# AI Consultants v5.1.0 - AI Expert Panel
+# AI Consultants v5.1.2 - AI Expert Panel
 
 **Coverage, not a single guess.** Convene a panel of AI "consultants" from different vendors for coding questions: they fan out in parallel and you get the *union* of what they collectively see — the risks, edge cases, and approaches a single model misses. Each consultant has a **configurable persona** that decorrelates its analysis.
 
@@ -86,7 +86,7 @@ Configuration (presets, strategies, features, personas, API keys) can be managed
 
 **API-only consultants**: GLM (The Methodologist), DeepSeek (The Code Specialist)
 
-**CLI/API Mode**: Gemini, Codex, Claude, Mistral, Qwen, Grok, and MiniMax can switch between CLI and API mode via `*_USE_API` environment variables. Grok is CLI-first on Grok Build with `grok-4.6` and falls back to the xAI API when the CLI cannot run. Gemini auto-selects API mode when `GEMINI_API_KEY` is set (no `agy` install needed) and the CLI otherwise. Mistral keeps separate API (`MISTRAL_MODEL`) and Vibe (`MISTRAL_CLI_MODEL`) identifiers.
+**CLI/API Mode**: Gemini, Codex, Claude, Mistral, Qwen, Grok, and MiniMax can switch between CLI and API mode via `*_USE_API` environment variables. Grok is CLI-first on Grok Build with `grok-4.7` and falls back to the xAI API when the CLI cannot run. Gemini auto-selects API mode when `GEMINI_API_KEY` is set (no `agy` install needed) and the CLI otherwise. Mistral keeps separate API (`MISTRAL_MODEL`) and Vibe (`MISTRAL_CLI_MODEL`) identifiers.
 
 Grok CLI OAuth defaults to `GROK_OAUTH_MODE=shared`: concurrent invocations
 share one runner-owned persistent `GROK_HOME` so the vendor auth lock can
@@ -103,10 +103,10 @@ the observed version is recorded only as response provenance.
 
 The `max_quality` preset enables all 10 consultants. It uses the separately
 smoke-tested K3-256k and MiniMax M3 targets, plus Qwen3.8-Max when an authenticated Token Plan transport is
-already configured. Gemini 3.7 Flash High is the exact-transport-smoked `agy`
-default for maximum, premium, and standard CLI tiers; Low serves economy. The
-Gemini API target remains 3.1 Pro until `gemini-3.7-flash` completes a separate
-API smoke. Claude Fable 5.1 is the default/premium/maximum target; the new
+already configured. Gemini 3.8 Flash High is the new `agy` target for maximum,
+premium, and standard CLI tiers; Low serves economy. The installed model inventory
+lists both; High/Low CLI smokes passed on 2026-09-28 with capability-probed identity.
+The Gemini API default remains 3.1 Pro; `gemini-3.8-flash` is an explicit API opt-in. Claude Fable 5.1 is the default/premium/maximum target; the new
 Mistral API IDs remain explicit opt-ins.
 Grok, GLM, and DeepSeek receive their
 highest accepted reasoning effort in this preset (`xhigh`, `max`, and `max`

package/docs/COST_RATES.md

@@ -1,4 +1,4 @@
-# Cost Rates - AI Consultants v5.1.0
+# Cost Rates - AI Consultants v5.1.2
 
 This page documents the per-token rates used by the cost tracking system.
 
@@ -23,8 +23,8 @@ models programmatically.
 
 | Model | Tier | Cost/1K | Consultant |
 |-------|------|---------|------------|
-| Gemini 3.7 Flash (High) | Maximum/Premium/Standard | $0.00075 | Gemini (verified agy CLI; promotional through 2026-12-31) |
-| Gemini 3.7 Flash (Low) | Economy | $0.00075 | Gemini (verified agy CLI; promotional through 2026-12-31) |
+| Gemini 3.8 Flash (High) | Maximum/Premium/Standard | $0.00075 | Gemini (CLI smoke passed; API-equivalent estimate; promotional through 2026-12-31) |
+| Gemini 3.8 Flash (Low) | Economy | $0.00075 | Gemini (CLI smoke passed; API-equivalent estimate; promotional through 2026-12-31) |
 | Gemini 3.1 Pro (High) | CLI opt-in | $0.002 | Gemini (agy CLI) |
 | gemini-3.1-pro-preview | API default | $0.00125 | Gemini API |
 | Gemini 3.6 Flash (High/Low) | Legacy CLI pins | $0.0015 | Gemini |
@@ -33,7 +33,8 @@ models programmatically.
 | gpt-5.6-terra | Standard | $0.002 | Codex |
 | gpt-5.6-luna | Economy | $0.0002 | Codex |
 | claude-fable-5-1 | Maximum/Premium | $0.010 | Claude |
-| claude-opus-5 | Standard | $0.005 | Claude |
+| claude-opus-5-5 | Standard | $0.004 | Claude |
+| claude-opus-5 | Legacy override | $0.005 | Claude |
 | claude-fable-5 | Legacy pins | $0.010 | Claude |
 | claude-sonnet-5 | Legacy pins | $0.003 | Claude |
 | claude-haiku-4-5 | Economy | $0.001 | Claude |
@@ -45,12 +46,14 @@ models programmatically.
 | composer-2.5 | Legacy | $0.0005 | Removed Cursor consultant |
 | composer-2 | Legacy | $0.0005 | Removed Cursor consultant |
 | gemini-3-flash | Legacy | $0.0005 | Removed Cursor consultant |
-| deepseek-v4-pro | Premium | $0.000435 | DeepSeek |
-| deepseek-v4-flash | Standard/Economy | $0.00014 | DeepSeek |
+| deepseek-v4-pro | Legacy explicit override | $0.000435 | DeepSeek |
+| deepseek-flash | All tiers | $0.0003 | DeepSeek (peak, cache miss) |
+| deepseek-v4-flash / deepseek-v4-flash-vision-exp | Retired aliases → V4.1 Flash | $0.0003 | DeepSeek (peak, cache miss) |
 | glm-5.3-flash | Maximum/Premium/Standard | Unpriced | GLM coding-plan endpoint |
 | glm-5.3 | Legacy pins | Unpriced | GLM coding-plan endpoint |
 | glm-4-flash | Economy | $0.001 | GLM |
-| grok-4.6 | Maximum/Premium | $0.002 | Grok |
+| grok-4.7 | Maximum/Premium | $0.002 | Grok |
+| grok-4.6 | Legacy override | $0.002 | Grok |
 | grok-4.5 | Standard/Economy | $0.002 | Grok |
 | qwen3.7-max | Premium | $0.0012 | Qwen3 |
 | qwen3.6-35b-a3b | Standard | $0.000163 | Qwen3 |
@@ -67,8 +70,8 @@ models programmatically.
 
 | Model | Tier | Cost/1K | Consultant |
 |-------|------|---------|------------|
-| Gemini 3.7 Flash (High) | Maximum/Premium/Standard | $0.00375 | Gemini (verified agy CLI; promotional through 2026-12-31) |
-| Gemini 3.7 Flash (Low) | Economy | $0.00375 | Gemini (verified agy CLI; promotional through 2026-12-31) |
+| Gemini 3.8 Flash (High) | Maximum/Premium/Standard | $0.00375 | Gemini (CLI smoke passed; API-equivalent estimate; promotional through 2026-12-31) |
+| Gemini 3.8 Flash (Low) | Economy | $0.00375 | Gemini (CLI smoke passed; API-equivalent estimate; promotional through 2026-12-31) |
 | Gemini 3.1 Pro (High) | CLI opt-in | $0.012 | Gemini (agy CLI) |
 | gemini-3.1-pro-preview | API default | $0.005 | Gemini API |
 | Gemini 3.6 Flash (High/Low) | Legacy CLI pins | $0.0075 | Gemini |
@@ -77,7 +80,8 @@ models programmatically.
 | gpt-5.6-terra | Standard | $0.012 | Codex |
 | gpt-5.6-luna | Economy | $0.0012 | Codex |
 | claude-fable-5-1 | Maximum/Premium | $0.050 | Claude |
-| claude-opus-5 | Standard | $0.025 | Claude |
+| claude-opus-5-5 | Standard | $0.020 | Claude |
+| claude-opus-5 | Legacy override | $0.025 | Claude |
 | claude-fable-5 | Legacy pins | $0.050 | Claude |
 | claude-sonnet-5 | Legacy pins | $0.015 | Claude |
 | claude-haiku-4-5 | Economy | $0.005 | Claude |
@@ -89,12 +93,14 @@ models programmatically.
 | composer-2.5 | Legacy | $0.0025 | Removed Cursor consultant |
 | composer-2 | Legacy | $0.0025 | Removed Cursor consultant |
 | gemini-3-flash | Legacy | $0.003 | Removed Cursor consultant |
-| deepseek-v4-pro | Premium | $0.00087 | DeepSeek |
-| deepseek-v4-flash | Standard/Economy | $0.00028 | DeepSeek |
+| deepseek-v4-pro | Legacy explicit override | $0.00087 | DeepSeek |
+| deepseek-flash | All tiers | $0.0012 | DeepSeek (peak) |
+| deepseek-v4-flash / deepseek-v4-flash-vision-exp | Retired aliases → V4.1 Flash | $0.0012 | DeepSeek (peak) |
 | glm-5.3-flash | Maximum/Premium/Standard | Unpriced | GLM coding-plan endpoint |
 | glm-5.3 | Legacy pins | Unpriced | GLM coding-plan endpoint |
 | glm-4-flash | Economy | $0.003 | GLM |
-| grok-4.6 | Maximum/Premium | $0.006 | Grok |
+| grok-4.7 | Maximum/Premium | $0.006 | Grok |
+| grok-4.6 | Legacy override | $0.006 | Grok |
 | grok-4.5 | Standard/Economy | $0.006 | Grok |
 | qwen3.7-max | Premium | $0.006 | Qwen3 |
 | qwen3.6-35b-a3b | Standard | $0.0009 | Qwen3 |
@@ -191,3 +197,37 @@ presenting them as free.
 Cost tracking requires:
 - `bc` - For floating point calculations
 - `jq` - For JSON parsing
+
+## DeepSeek V4.1 Flash pricing
+
+As of 2026-09-10, the official API name `deepseek-flash` selects V4.1 Flash.
+The estimator uses peak cache-miss rates: $0.30/M input and $1.20/M output.
+Off-peak rates are $0.15/M input and $0.60/M output; cached input is $0.006/M
+peak or $0.003/M off-peak. Peak hours are weekdays 01:00–04:00 and 06:00–10:00
+UTC. The estimate deliberately excludes these discounts; it is not an invoice.
+Legacy Flash and Flash Vision Exp aliases are now served and billed as V4.1 Flash.
+Other legacy catalog entries remain historical estimates, not newly verified quotes.
+Source: [DeepSeek models and pricing](https://api-docs.deepseek.com/quick_start/pricing/).
+
+## Grok 4.7 and Opus 5.5
+
+Grok 4.7 uses $2/M input and $6/M output at standard API rates; above 200K
+input tokens the estimator applies $4/M and $12/M to the whole request. It
+excludes cached-input discounts and regional, priority and Grok Build Fast
+adjustments. CLI subscription usage is not an API invoice. Opus 5.5 uses
+$4/M input and $20/M output; measured Claude provider cost takes precedence
+when present. Otherwise cache/service adjustments are excluded from estimates.
+
+Sources checked 2026-09-26: [xAI release notes](https://docs.x.ai/developers/release-notes)
+and [Opus 5.5](https://platform.claude.com/docs/en/models/opus-5-5/overview).
+
+## Gemini 3.8 Flash
+
+CLI High/Medium/Low and API ID `gemini-3.8-flash` are priced at the promotional
+API-equivalent rate: $0.75/M input, $3.75/M output through 2026-12-31. Standard
+rates from 2027-01-01 are $1.50/M input and $7.50/M output; the static catalog
+must be refreshed at expiry. Cache, service-tier and subscription adjustments
+are excluded: a CLI estimate is not a subscription invoice. Legacy model rates
+remain in the catalog. API Pro 3.1 is still the default for API mode.
+Source: [Google latest-model guide](https://ai.google.dev/gemini-api/docs/latest-model),
+checked 2026-09-27.

package/docs/SETUP.md

@@ -850,3 +850,84 @@ argument list. Hidden flags such as `--no-memory` are still required in the
 actual request even when omitted from public help. Unsupported mandatory flags
 block dispatch; the optional `--no-auto-update` guard is used when its parser
 probe succeeds, including on CLIs that hide it from help.
+
+### DeepSeek V4.1 Flash
+
+DeepSeek uses the HTTP API with `DEEPSEEK_MODEL=deepseek-flash` in every tier.
+This is the official API alias for DeepSeek-V4.1-Flash as of 2026-09-10, not a
+version identity inferred from the requested name. Provider-reported model
+metadata is retained unchanged. `max_quality` continues to request `max` effort;
+other tiers leave effort to the provider unless explicitly configured. The
+provider defaults to thinking mode at high effort. Existing text-only Chat
+Completions requests and the 16,384-token completion budget remain compatible.
+
+`configure` migrates exact unpinned `deepseek-v4-pro` and `deepseek-v4-flash`
+values; `# ai-consultants:pin`, environment overrides and `--set` preserve explicit
+choices. Old Flash aliases are still accepted but route to V4.1 Flash upstream.
+An explicit pin preserves the requested ID, not a guarantee against provider
+routing changes. Peak cache-miss costs are estimates, excluding cache/off-peak
+discounts. An authorized one-attempt API smoke on 2026-09-10 passed with a structured
+response and provider-reported `deepseek-flash`; no version identity was invented
+from the alias. No CLI transport is introduced.
+
+Sources: [models/pricing](https://api-docs.deepseek.com/quick_start/pricing/) and
+[thinking mode](https://api-docs.deepseek.com/guides/thinking_mode/).
+
+### Grok 4.7 and Claude Opus 5.5
+
+Grok defaults to `grok-4.7` for premium/maximum; standard/economy stay on
+`grok-4.5`. Configure migrates exact unpinned persisted `grok-4.5`/`grok-4.6`
+defaults. Pins, environment overrides, `--set` and unrelated IDs survive.
+The existing capability/model inventory probes, strict sandbox and OAuth
+isolation still apply; the Grok 4.7 Fast variant is not implicitly selected.
+
+Claude standard/medium selects `claude-opus-5-5`; default/premium/maximum
+remain Fable 5.1 and economy remains Haiku 4.5. Select Opus directly with
+`configure --set CLAUDE_MODEL=claude-opus-5-5`. The historical unpinned Claude
+default migration to Fable is unchanged; explicit old Opus pins stay pinned.
+
+Opus 5.5 has always-on adaptive thinking, including empty thinking display
+blocks. Our tool-free single-turn API request omits thinking overrides, sampling
+parameters and forced tool choice; parsing selects visible text by block type.
+The API therefore uses the provider's default effort (medium), with the existing
+16384 total-output budget. The CLI retains explicit `CLAUDE_REASONING_EFFORT`
+control. Content identity and terminal success remain separate checks.
+
+These additions have offline regression coverage. Grok 4.7 and Opus 5.5 CLI
+smokes passed on 2026-09-28. New API transports remain separately unverified. Frozen
+P1.6 benchmark IDs/efforts/preregistration are not changed by this update.
+See the [Opus migration guide](https://platform.claude.com/docs/en/models/opus-5-5/migration-guide)
+and [Grok 4.7](https://docs.x.ai/developers/grok-4-7).
+
+### Gemini 3.8 Flash
+
+The installed `agy models` inventory on 2026-09-27 lists Gemini 3.8 Flash
+(High), (Medium) and (Low). CLI default/maximum/premium/standard now select High;
+economy selects Low. Medium remains an explicit `GEMINI_MODEL` choice. Inventory
+evidence is capability-probed, not provider attestation of a generated answer.
+
+Configure migrates unpinned historical Pro 3.1 High and Flash 3.7 High/Low
+defaults, preserving the High/Low choice. Pins, environment/--set overrides and
+unrelated models are preserved. API mode remains on `gemini-3.1-pro-preview`.
+Select `GEMINI_API_MODEL=gemini-3.8-flash` explicitly to use Flash via API; its
+native thinking levels are low/medium/high. Other explicit efforts fail before
+dispatch. With no effort setting, the provider's API default is medium.
+
+Inventory discovery performed no inference. Separately authorized High/Low CLI
+smokes passed on 2026-09-28, using the named variant without --effort. This is
+capability-probed provenance, not provider content-model attestation. Prices are promotional estimates
+through 2026-12-31; see [Google's model guide](https://ai.google.dev/gemini-api/docs/latest-model).
+The frozen P1.6 inputs remain unchanged.
+
+For Google GenerateContent accounting, output tokens include both
+`candidatesTokenCount` and `thoughtsTokenCount`. Input uses `promptTokenCount`,
+which already includes cached content. This avoids dropping billed thinking
+or adding cache tokens twice; it applies to the existing Google API path too.
+See [UsageMetadata](https://ai.google.dev/api/generate-content).
+
+For named Gemini 3.8 Flash High/Medium/Low CLI variants, leave
+`GEMINI_REASONING_EFFORT` unset: the level is selected by `GEMINI_MODEL`, and
+`agy` rejects an additional `--effort` for these variants. The CLI's general
+help lists effort values for other supported model selections; it does not
+establish support for every model. Native API thinking remains separately
+configured with low/medium/high.

package/docs/cost_rates.json

@@ -1,7 +1,49 @@
 {
-  "version": "5.1.0",
-  "updated": "2026-08-23",
+  "version": "5.1.2",
+  "updated": "2026-09-28",
   "models": {
+    "gemini-3.8-flash": {
+      "input": 0.00075,
+      "output": 0.00375,
+      "notes": "Promotional API-equivalent estimate through 2026-12-31; standard from 2027-01-01: 0.0015/0.0075 per 1K. Cache/service adjustments excluded; not a CLI invoice."
+    },
+    "Gemini 3.8 Flash (High)": {
+      "input": 0.00075,
+      "output": 0.00375,
+      "notes": "Promotional API-equivalent estimate through 2026-12-31; standard from 2027-01-01: 0.0015/0.0075 per 1K. Cache/service adjustments excluded; not a CLI invoice."
+    },
+    "Gemini 3.8 Flash (Medium)": {
+      "input": 0.00075,
+      "output": 0.00375,
+      "notes": "Promotional API-equivalent estimate through 2026-12-31; standard from 2027-01-01: 0.0015/0.0075 per 1K. Cache/service adjustments excluded; not a CLI invoice."
+    },
+    "Gemini 3.8 Flash (Low)": {
+      "input": 0.00075,
+      "output": 0.00375,
+      "notes": "Promotional API-equivalent estimate through 2026-12-31; standard from 2027-01-01: 0.0015/0.0075 per 1K. Cache/service adjustments excluded; not a CLI invoice."
+    },
+    "claude-opus-5-5": {
+      "input": 0.004,
+      "output": 0.020,
+      "notes": "Standard API list-rate estimate; cache/service adjustments excluded. Verified 2026-09-26."
+    },
+    "grok-4.7": {
+      "input": 0.002,
+      "output": 0.006,
+      "notes": "Standard API estimate; above 200K input both rates double for the request. Cache, regional, priority and Fast rates excluded. Verified 2026-09-26."
+    },
+    "deepseek-flash": {
+      "input": 0.0003,
+      "output": 0.0012,
+      "provider": "DeepSeek",
+      "notes": "DeepSeek-V4.1-Flash, verified 2026-09-10. Peak cache-miss estimate; off-peak and cache-hit discounts excluded. See https://api-docs.deepseek.com/quick_start/pricing/"
+    },
+    "deepseek-v4-flash-vision-exp": {
+      "input": 0.0003,
+      "output": 0.0012,
+      "provider": "DeepSeek",
+      "notes": "Retired alias served by V4.1 Flash; peak cache-miss estimate."
+    },
     "_comment_units": "All rates are USD per 1,000 tokens (per-1K). lib/costs.sh computes cost as (token_count / 1000) * rate, so every value here MUST be a per-1K rate = (provider $/1M price) / 1000. Do NOT paste $/MTok figures directly.",
     "_comment_premium": "Premium tier models - Sep 2026",
     "claude-fable-5-1": {
@@ -380,8 +422,8 @@
       "output": 0.003
     },
     "deepseek-v4-flash": {
-      "input": 0.00014,
-      "output": 0.00028
+      "input": 0.0003,
+      "output": 0.0012
     },
     "glm-5.2": {
       "input": 0.00098,
@@ -431,12 +473,12 @@
   "consultant_fallbacks": {
     "_comment": "Premium defaults - Updated Sep 2026",
     "claude": "claude-fable-5-1",
-    "gemini": "Gemini 3.7 Flash (High)",
+    "gemini": "Gemini 3.8 Flash (High)",
     "codex": "gpt-6-astra",
     "mistral": "mistral-large-3",
-    "deepseek": "deepseek-v4-pro",
+    "deepseek": "deepseek-flash",
     "glm": "glm-5.3-flash",
-    "grok": "grok-4.6",
+    "grok": "grok-4.7",
     "qwen3": "qwen3.7-max",
     "kimi": "kimi-code/k3",
     "minimax": "minimax-m2.7"
@@ -445,34 +487,34 @@
     "_comment": "Model tiers for apply_model_tier() function",
     "maximum": {
       "claude": "claude-fable-5-1",
-      "gemini": "Gemini 3.7 Flash (High)",
+      "gemini": "Gemini 3.8 Flash (High)",
       "codex": "gpt-6-astra",
       "mistral": "mistral-medium-3.5",
-      "deepseek": "deepseek-v4-pro",
+      "deepseek": "deepseek-flash",
       "glm": "glm-5.3-flash",
-      "grok": "grok-4.6",
+      "grok": "grok-4.7",
       "qwen3": "qwen3.8-max",
       "kimi": "kimi-code/k3-256k",
       "minimax": "MiniMax-M3"
     },
     "premium": {
       "claude": "claude-fable-5-1",
-      "gemini": "Gemini 3.7 Flash (High)",
+      "gemini": "Gemini 3.8 Flash (High)",
       "codex": "gpt-6-astra",
       "mistral": "mistral-medium-3.5",
-      "deepseek": "deepseek-v4-pro",
+      "deepseek": "deepseek-flash",
       "glm": "glm-5.3-flash",
-      "grok": "grok-4.6",
+      "grok": "grok-4.7",
       "qwen3": "qwen3.7-max",
       "kimi": "kimi-code/k3",
       "minimax": "MiniMax-M2.7"
     },
     "standard": {
-      "claude": "claude-opus-5",
-      "gemini": "Gemini 3.7 Flash (High)",
+      "claude": "claude-opus-5-5",
+      "gemini": "Gemini 3.8 Flash (High)",
       "codex": "gpt-5.6-terra",
       "mistral": "mistral-medium-3.5",
-      "deepseek": "deepseek-v4-flash",
+      "deepseek": "deepseek-flash",
       "glm": "glm-5.3-flash",
       "grok": "grok-4.5",
       "qwen3": "qwen3.6-35b-a3b",
@@ -481,10 +523,10 @@
     },
     "economy": {
       "claude": "claude-haiku-4-5",
-      "gemini": "Gemini 3.7 Flash (Low)",
+      "gemini": "Gemini 3.8 Flash (Low)",
       "codex": "gpt-5.6-luna",
       "mistral": "devstral-small-2",
-      "deepseek": "deepseek-v4-flash",
+      "deepseek": "deepseek-flash",
       "glm": "glm-4-flash",
       "grok": "grok-4.5",
       "qwen3": "qwen3-32b",
@@ -495,10 +537,10 @@
   "transport_model_tiers": {
     "_comment": "Transport-specific namespaces used by Gemini and Mistral; the legacy model_tiers table reflects their CLI values.",
     "gemini": {
-      "maximum": {"cli": "Gemini 3.7 Flash (High)", "api": "gemini-3.1-pro-preview"},
-      "premium": {"cli": "Gemini 3.7 Flash (High)", "api": "gemini-3.1-pro-preview"},
-      "standard": {"cli": "Gemini 3.7 Flash (High)", "api": "gemini-3.1-pro-preview"},
-      "economy": {"cli": "Gemini 3.7 Flash (Low)", "api": "gemini-3.1-pro-preview"}
+      "maximum": {"cli": "Gemini 3.8 Flash (High)", "api": "gemini-3.1-pro-preview"},
+      "premium": {"cli": "Gemini 3.8 Flash (High)", "api": "gemini-3.1-pro-preview"},
+      "standard": {"cli": "Gemini 3.8 Flash (High)", "api": "gemini-3.1-pro-preview"},
+      "economy": {"cli": "Gemini 3.8 Flash (Low)", "api": "gemini-3.1-pro-preview"}
     },
     "mistral": {
       "maximum": {"cli": "mistral-medium-3.5", "api": "mistral-large-3"},
@@ -508,8 +550,8 @@
     }
   },
   "opt_in_targets": {
-    "gemini_cli": ["Gemini 3.1 Pro (High)", "Gemini 3.6 Flash (High)", "Gemini 3.6 Flash (Low)"],
-    "gemini_api": ["gemini-3.7-flash"],
+    "gemini_cli": ["Gemini 3.8 Flash (Medium)", "Gemini 3.7 Flash (High)", "Gemini 3.7 Flash (Low)", "Gemini 3.1 Pro (High)", "Gemini 3.6 Flash (High)", "Gemini 3.6 Flash (Low)"],
+    "gemini_api": ["gemini-3.7-flash", "gemini-3.8-flash"],
     "mistral_api": ["mistral-medium-3-5", "mistral-large-2512", "mistral-small-2603"],
     "claude": []
   },

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "ai-consultants",
-  "version": "5.1.0",
+  "version": "5.1.2",
   "description": "A cross-vendor panel of up to 10 AI models that returns the coverage union of their distinct perspectives",
   "bin": {
     "ai-consultants": "bin/ai-consultants"

package/references/configuration.md

@@ -59,7 +59,7 @@ Model defaults carry an `# ai-consultants:default` marker. On rewrite,
 `CLAUDE_MODEL=claude-opus-4-8` and `claude-opus-5` generated defaults to
 Fable 5.1. An explicit model override supplied with `--set` is stored with
 `# ai-consultants:pin`, so `--set CLAUDE_MODEL=claude-opus-5` keeps the
-lower-cost standard model intentionally.
+legacy Opus 5 model intentionally. Standard/medium now select Opus 5.5.
 The exhaustive parameter contract is [`.env.example`](../.env.example); the
 configurator derives its accepted keys from that template, while
 [`scripts/config.sh`](../scripts/config.sh) remains the runtime source of truth.
@@ -240,11 +240,11 @@ ENABLE_DEEPSEEK=false
 ### Model Overrides
 
 ```bash
-GEMINI_MODEL=Gemini 3.7 Flash (High) # verified agy CLI default; API mode uses GEMINI_API_MODEL
+GEMINI_MODEL="Gemini 3.8 Flash (High)" # CLI smoke passed; capability-probed identity; API uses GEMINI_API_MODEL
 CODEX_MODEL=gpt-6-astra
 CODEX_API_MAX_TOKENS=16384
 CLAUDE_MODEL=claude-fable-5-1
-# Lower-cost standard override: CLAUDE_MODEL=claude-opus-5
+# Lower-cost standard override: CLAUDE_MODEL=claude-opus-5-5
 CLAUDE_API_MAX_TOKENS=16384  # API only: adaptive thinking + visible output
 MISTRAL_MODEL=mistral-large-3
 MISTRAL_CLI_MODEL=mistral-medium-3.5
@@ -257,12 +257,12 @@ QWEN3_API_MAX_TOKENS=16384
 GLM_MODEL=glm-5.3-flash
 GLM_API_MAX_TOKENS=16384
 GLM_REASONING_EFFORT=       # max_quality sets max
-GROK_MODEL=grok-4.6
+GROK_MODEL=grok-4.7
 GROK_MAX_TURNS=4
 GROK_API_MAX_TOKENS=4096
 GROK_REASONING_EFFORT=      # max_quality sets xhigh (Grok Build maximum)
 GROK_OAUTH_MODE=shared      # shared (parallel default) | serialized (diagnostic)
-DEEPSEEK_MODEL=deepseek-v4-pro
+DEEPSEEK_MODEL=deepseek-flash  # DeepSeek-V4.1-Flash API alias
 DEEPSEEK_API_MAX_TOKENS=16384
 DEEPSEEK_MAX_QUALITY_TIMEOUT=600
 DEEPSEEK_REASONING_EFFORT=  # max_quality sets max

package/scripts/config.sh

@@ -128,7 +128,7 @@ LAUNCH_STAGGER_MAX_SECONDS="${LAUNCH_STAGGER_MAX_SECONDS:-2}"
 
 # CLI mode uses the Antigravity CLI (`agy`), successor to the deprecated
 # Gemini CLI (transitioned 2026-06-18). Models are passed by display name.
-GEMINI_MODEL="${GEMINI_MODEL:-Gemini 3.7 Flash (High)}"
+GEMINI_MODEL="${GEMINI_MODEL:-Gemini 3.8 Flash (High)}"
 GEMINI_TIMEOUT_SECONDS="${GEMINI_TIMEOUT:-180}"
 GEMINI_CMD="${GEMINI_CMD:-agy}"
 # API mode (GEMINI_USE_API=true) talks to the Google AI generativelanguage
@@ -213,7 +213,7 @@ GLM_FORMAT="${GLM_FORMAT:-openai}"
 # =============================================================================
 
 GROK_CMD="${GROK_CMD:-grok}"
-GROK_MODEL="${GROK_MODEL:-grok-4.6}"
+GROK_MODEL="${GROK_MODEL:-grok-4.7}"
 GROK_TIMEOUT_SECONDS="${GROK_TIMEOUT:-180}"
 GROK_API_MAX_TOKENS="${GROK_API_MAX_TOKENS:-4096}"
 GROK_MAX_TURNS="${GROK_MAX_TURNS:-4}"
@@ -243,7 +243,7 @@ export GROK_USE_API GROK_OAUTH_MODE
 # DEEPSEEK CONFIGURATION - The Methodologist (API-based)
 # =============================================================================
 
-DEEPSEEK_MODEL="${DEEPSEEK_MODEL:-deepseek-v4-pro}"
+DEEPSEEK_MODEL="${DEEPSEEK_MODEL:-deepseek-flash}"
 if [[ -n "${DEEPSEEK_TIMEOUT+x}" ]]; then
     DEEPSEEK_TIMEOUT_SECONDS="$DEEPSEEK_TIMEOUT"
 elif [[ -z "${DEEPSEEK_TIMEOUT_SECONDS+x}" ]]; then
@@ -619,12 +619,12 @@ get_model_for_tier() {
         maximum|max_quality|max-quality)
             case "$consultant" in
                 claude)   echo "claude-fable-5-1" ;;
-                gemini)   [[ "$transport" == "api" ]] && echo "gemini-3.1-pro-preview" || echo "Gemini 3.7 Flash (High)" ;;
+                gemini)   [[ "$transport" == "api" ]] && echo "gemini-3.1-pro-preview" || echo "Gemini 3.8 Flash (High)" ;;
                 codex)    echo "gpt-6-astra" ;;
                 mistral)  [[ "$transport" == "api" ]] && echo "mistral-large-3" || echo "mistral-medium-3.5" ;;
-                deepseek) echo "deepseek-v4-pro" ;;
+                deepseek) echo "deepseek-flash" ;;
                 glm)      echo "glm-5.3-flash" ;;
-                grok)     echo "grok-4.6" ;;
+                grok)     echo "grok-4.7" ;;
                 qwen3)    echo "qwen3.8-max" ;;
                 kimi)     echo "kimi-code/k3-256k" ;;
                 minimax)  echo "MiniMax-M3" ;;
@@ -634,12 +634,12 @@ get_model_for_tier() {
         premium|max|best)
             case "$consultant" in
                 claude)   echo "claude-fable-5-1" ;;
-                gemini)   [[ "$transport" == "api" ]] && echo "gemini-3.1-pro-preview" || echo "Gemini 3.7 Flash (High)" ;;
+                gemini)   [[ "$transport" == "api" ]] && echo "gemini-3.1-pro-preview" || echo "Gemini 3.8 Flash (High)" ;;
                 codex)    echo "gpt-6-astra" ;;
                 mistral)  [[ "$transport" == "api" ]] && echo "mistral-large-3" || echo "mistral-medium-3.5" ;;
-                deepseek) echo "deepseek-v4-pro" ;;
+                deepseek) echo "deepseek-flash" ;;
                 glm)      echo "glm-5.3-flash" ;;
-                grok)     echo "grok-4.6" ;;
+                grok)     echo "grok-4.7" ;;
                 qwen3)    echo "qwen3.7-max" ;;
                 kimi)     echo "kimi-code/k3" ;;
                 minimax)  echo "MiniMax-M2.7" ;;
@@ -648,11 +648,11 @@ get_model_for_tier() {
             ;;
         standard|medium|balanced)
             case "$consultant" in
-                claude)   echo "claude-opus-5" ;;
-                gemini)   [[ "$transport" == "api" ]] && echo "gemini-3.1-pro-preview" || echo "Gemini 3.7 Flash (High)" ;;
+                claude)   echo "claude-opus-5-5" ;;
+                gemini)   [[ "$transport" == "api" ]] && echo "gemini-3.1-pro-preview" || echo "Gemini 3.8 Flash (High)" ;;
                 codex)    echo "gpt-5.6-terra" ;;
                 mistral)  [[ "$transport" == "api" ]] && echo "mistral-large-3" || echo "mistral-medium-3.5" ;;
-                deepseek) echo "deepseek-v4-flash" ;;
+                deepseek) echo "deepseek-flash" ;;
                 glm)      echo "glm-5.3-flash" ;;  # Same as premium (no mid-tier GLM)
                 grok)     echo "grok-4.5" ;;
                 qwen3)    echo "qwen3.6-35b-a3b" ;;  # Open-weight MoE (35B total, 3B active)
@@ -664,10 +664,10 @@ get_model_for_tier() {
         economy|fast|quick)
             case "$consultant" in
                 claude)   echo "claude-haiku-4-5" ;;
-                gemini)   [[ "$transport" == "api" ]] && echo "gemini-3.1-pro-preview" || echo "Gemini 3.7 Flash (Low)" ;;
+                gemini)   [[ "$transport" == "api" ]] && echo "gemini-3.1-pro-preview" || echo "Gemini 3.8 Flash (Low)" ;;
                 codex)    echo "gpt-5.6-luna" ;;
                 mistral)  [[ "$transport" == "api" ]] && echo "mistral-large-3" || echo "devstral-small-2" ;;
-                deepseek) echo "deepseek-v4-flash" ;;
+                deepseek) echo "deepseek-flash" ;;
                 glm)      echo "glm-4-flash" ;;
                 grok)     echo "grok-4.5" ;;
                 qwen3)    echo "qwen3-32b" ;;
@@ -715,7 +715,7 @@ apply_model_tier() {
 
     # Gemini and Mistral intentionally carry both transport-specific IDs so a
     # later transport choice cannot send a CLI display name to an HTTP API (or
-    # an API slug to a CLI). Gemini 3.7 uses one API ID plus a thinking level.
+    # an API slug to a CLI). Gemini 3.7/3.8 Flash use one API ID plus a thinking level.
     export GEMINI_MODEL="$(get_model_for_tier gemini "$tier" cli)"
     export GEMINI_API_MODEL="$(get_model_for_tier gemini "$tier" api)"
     export MISTRAL_CLI_MODEL="$(get_model_for_tier mistral "$tier" cli)"
@@ -839,8 +839,8 @@ apply_model_tier() {
             ;;
     esac
 
-    # Gemini 3.7 is promoted only on the exact agy transport that completed a
-    # live adapter smoke. API tiers remain on the separately proven 3.1 Pro ID;
+    # Gemini 3.8 CLI availability is inventory-probed; new live smokes are
+    # required separately. API tiers remain on the proven 3.1 Pro ID;
     # neither transport silently substitutes the other's model namespace.
 
     case "$tier" in
@@ -1035,4 +1035,4 @@ list_presets() {
 # VERSION
 # =============================================================================
 
-AI_CONSULTANTS_VERSION="5.1.0"
+AI_CONSULTANTS_VERSION="5.1.2"

package/scripts/configure.sh

@@ -133,6 +133,8 @@ chmod 600 "$WORK_FILE"
 
 read_value() {
     local file="$1" wanted="$2" line key value found="" seen=false
+    local double_quoted='^"(([^"\\]|\\.)*)"[[:space:]]*(#.*)?$'
+    local single_quoted="^'([^']*)'[[:space:]]*(#.*)?$"
     [[ -r "$file" ]] || return 1
     while IFS= read -r line || [[ -n "$line" ]]; do
         line="${line%$'\r'}"
@@ -147,8 +149,8 @@ read_value() {
         key="${key//[[:space:]]/}"
         [[ "$key" == "$wanted" ]] || continue
         value="${line#*=}"
-        if [[ "$value" =~ ^\".*\"$ ]] || [[ "$value" =~ ^\'.*\'$ ]]; then
-            value="${value:1:${#value}-2}"
+        if [[ "$value" =~ $double_quoted ]] || [[ "$value" =~ $single_quoted ]]; then
+            value="${BASH_REMATCH[1]}"
         else
             value=$(printf '%s' "$value" | sed -E 's/[[:space:]]+#.*$//; s/[[:space:]]+$//')
         fi
@@ -314,11 +316,20 @@ esac
 # Migrate only the exact catalog defaults previously generated by this tool.
 # A # ai-consultants:pin marker is the durable escape hatch for users who want
 # to keep an older target or a nonstandard binary name.
-if [[ "$(read_value "$WORK_FILE" GEMINI_MODEL 2>/dev/null || true)" == "Gemini 3.1 Pro (High)" ]] \
-    && ! has_value_marker "$WORK_FILE" GEMINI_MODEL "$PIN_MARKER"; then
-    set_value GEMINI_MODEL "Gemini 3.7 Flash (High) $DEFAULT_MARKER"
-    echo "Migrated managed Gemini CLI default: Gemini 3.1 Pro (High) -> Gemini 3.7 Flash (High)" >&2
-fi
+case "$(read_value "$WORK_FILE" GEMINI_MODEL 2>/dev/null || true)" in
+    'Gemini 3.1 Pro (High)'|'Gemini 3.7 Flash (High)')
+        if ! has_value_marker "$WORK_FILE" GEMINI_MODEL "$PIN_MARKER"; then
+            set_value GEMINI_MODEL "Gemini 3.8 Flash (High) $DEFAULT_MARKER"
+            echo "Migrated managed Gemini CLI default to Gemini 3.8 Flash (High)" >&2
+        fi
+        ;;
+    'Gemini 3.7 Flash (Low)')
+        if ! has_value_marker "$WORK_FILE" GEMINI_MODEL "$PIN_MARKER"; then
+            set_value GEMINI_MODEL "Gemini 3.8 Flash (Low) $DEFAULT_MARKER"
+            echo "Migrated managed Gemini CLI economy default to Gemini 3.8 Flash (Low)" >&2
+        fi
+        ;;
+esac
 
 case "$(read_value "$WORK_FILE" GLM_MODEL 2>/dev/null || true)" in
     glm-5.2|glm-5.3)
@@ -329,11 +340,25 @@ case "$(read_value "$WORK_FILE" GLM_MODEL 2>/dev/null || true)" in
         ;;
 esac
 
-if [[ "$(read_value "$WORK_FILE" GROK_MODEL 2>/dev/null || true)" == "grok-4.5" ]] \
-    && ! has_value_marker "$WORK_FILE" GROK_MODEL "$PIN_MARKER"; then
-    set_value GROK_MODEL "grok-4.6 $DEFAULT_MARKER"
-    echo "Migrated managed Grok default: grok-4.5 -> grok-4.6" >&2
-fi
+case "$(read_value "$WORK_FILE" GROK_MODEL 2>/dev/null || true)" in
+    grok-4.5|grok-4.6)
+        if ! has_value_marker "$WORK_FILE" GROK_MODEL "$PIN_MARKER"; then
+            set_value GROK_MODEL "grok-4.7 $DEFAULT_MARKER"
+            echo "Migrated managed Grok default to grok-4.7" >&2
+        fi
+        ;;
+esac
+
+# Refresh only exact unpinned historical DeepSeek defaults. Environment and
+# --set selections carry pin markers and remain user-owned.
+case "$(read_value "$WORK_FILE" DEEPSEEK_MODEL 2>/dev/null || true)" in
+    deepseek-v4-pro|deepseek-v4-flash)
+        if ! has_value_marker "$WORK_FILE" DEEPSEEK_MODEL "$PIN_MARKER"; then
+            set_value DEEPSEEK_MODEL "deepseek-flash $DEFAULT_MARKER"
+            echo "Migrated managed DeepSeek default to deepseek-flash (V4.1 Flash)" >&2
+        fi
+        ;;
+esac
 
 configure_cli_only() {
     local enable_key="$1" cmd_key="$2" cmd enabled=false

package/scripts/lib/api.sh

@@ -345,7 +345,11 @@ extract_token_usage() {
             ;;
         google_ai)
             input_tokens=$(echo "$response" | jq -r '.usageMetadata.promptTokenCount // 0' 2>/dev/null)
-            output_tokens=$(echo "$response" | jq -r '.usageMetadata.candidatesTokenCount // 0' 2>/dev/null)
+            # Google bills candidates plus thoughts as output. Cached input is
+            # already included in promptTokenCount and must not be added again.
+            output_tokens=$(echo "$response" | jq -r '
+                def token: if type == "number" and . >= 0 and floor == . then . else 0 end;
+                .usageMetadata | ((.candidatesTokenCount | token) + (.thoughtsTokenCount | token))' 2>/dev/null)
             ;;
         *)  # openai format (default)
             input_tokens=$(echo "$response" | jq -r '.usage.prompt_tokens // 0' 2>/dev/null)
@@ -381,7 +385,11 @@ extract_token_split() {
             ;;
         google_ai)
             input_tokens=$(echo "$response" | jq -r '.usageMetadata.promptTokenCount // 0' 2>/dev/null)
-            output_tokens=$(echo "$response" | jq -r '.usageMetadata.candidatesTokenCount // 0' 2>/dev/null)
+            # Google bills candidates plus thoughts as output. Cached input is
+            # already included in promptTokenCount and must not be added again.
+            output_tokens=$(echo "$response" | jq -r '
+                def token: if type == "number" and . >= 0 and floor == . then . else 0 end;
+                .usageMetadata | ((.candidatesTokenCount | token) + (.thoughtsTokenCount | token))' 2>/dev/null)
             ;;
         *)  # openai format (default)
             input_tokens=$(echo "$response" | jq -r '.usage.prompt_tokens // 0' 2>/dev/null)

package/scripts/lib/api_query.sh

@@ -88,24 +88,24 @@ run_api_mode_query() {
 
     case "$api_format" in
         google_ai)
-            # Gemini 3.7 exposes low|medium|high thinking levels on the native
+            # Gemini 3.7/3.8 Flash expose low|medium|high thinking levels on the native
             # Google AI wire. The shared syntax validator is intentionally
             # broader for OpenAI-compatible providers, so fail before spending
             # a request when Google receives an unsupported level.
             local google_thinking_level=""
             if [[ -n "$effort" ]]; then
                 case "$model" in
-                    gemini-3.7-flash*)
+                    gemini-3.7-flash*|gemini-3.8-flash)
                         case "$effort" in
                             low|medium|high) google_thinking_level="$effort" ;;
                             *)
-                                log_error "[$consultant_name] $effort_var=$effort is unsupported for Gemini 3.7 on the 'google_ai' wire format (expected low|medium|high)"
+                                log_error "[$consultant_name] $effort_var=$effort is unsupported for $model on the 'google_ai' wire format (expected low|medium|high)"
                                 return 1
                                 ;;
                         esac
                         ;;
                     *)
-                        log_warn "[$consultant_name] $effort_var is only transported to Gemini 3.7 on the 'google_ai' wire format and is ignored for $model."
+                        log_warn "[$consultant_name] $effort_var is only transported to Gemini 3.7/3.8 Flash on the 'google_ai' wire format and is ignored for $model."
                         ;;
                 esac
             fi

package/scripts/lib/costs.sh

@@ -133,9 +133,13 @@ get_input_cost_per_1k() {
 
     # Fallback to hardcoded rates for backwards compatibility
     case "$model" in
+        gemini-3.8-flash|"gemini 3.8 flash (high)"|"gemini 3.8 flash (medium)"|"gemini 3.8 flash (low)") echo "0.00075" ;;
         gemini-2.5-pro)   echo "0.00125" ;;
         gemini-2.5-flash) echo "0.000075" ;;
         gemini-2.0-flash) echo "0.0001" ;;
+        deepseek-flash|deepseek-v4-flash|deepseek-v4-flash-vision-exp) echo "0.0003" ;;
+        claude-opus-5-5) echo "0.004" ;;
+        grok-4.7) echo "0.002" ;;
         gpt-6-astra|claude-fable-5-1) echo "0.01" ;;
         gpt-4)            echo "0.03" ;;
         gpt-4-turbo)      echo "0.01" ;;
@@ -180,9 +184,13 @@ get_output_cost_per_1k() {
 
     # Fallback to hardcoded rates for backwards compatibility
     case "$model" in
+        gemini-3.8-flash|"gemini 3.8 flash (high)"|"gemini 3.8 flash (medium)"|"gemini 3.8 flash (low)") echo "0.00375" ;;
         gemini-2.5-pro)   echo "0.005" ;;
         gemini-2.5-flash) echo "0.0003" ;;
         gemini-2.0-flash) echo "0.0004" ;;
+        deepseek-flash|deepseek-v4-flash|deepseek-v4-flash-vision-exp) echo "0.0012" ;;
+        claude-opus-5-5) echo "0.020" ;;
+        grok-4.7) echo "0.006" ;;
         gpt-6-astra|claude-fable-5-1) echo "0.05" ;;
         gpt-4)            echo "0.06" ;;
         gpt-4-turbo)      echo "0.03" ;;
@@ -236,6 +244,10 @@ estimate_query_cost() {
         input_rate=$(echo "scale=6; $input_rate * 2" | bc)
         output_rate=$(echo "scale=6; $output_rate * 1.5" | bc)
     fi
+    if [[ "$model" == grok-4.7 && "$input_tokens" -gt 200000 ]]; then
+        input_rate=$(echo "scale=6; $input_rate * 2" | bc)
+        output_rate=$(echo "scale=6; $output_rate * 2" | bc)
+    fi
     # Calculate cost
     local input_cost output_cost total_cost
     input_cost=$(echo "scale=6; $input_tokens / 1000 * $input_rate" | bc)
@@ -401,7 +413,7 @@ format_cost_caveats() {
     local responses_dir="${1:-}"
     [[ -d "$responses_dir" ]] || return 0
 
-    local estimated=0 unknown=0 priced=0 f astra_estimate=0
+    local estimated=0 unknown=0 priced=0 f astra_estimate=0 deepseek_flash_estimate=0 grok_47_estimate=0 opus_55_estimate=0 gemini_38_estimate=0
     local unpriced_models=()
     while IFS= read -r f; do
         [[ -n "$f" ]] || continue
@@ -423,6 +435,24 @@ format_cost_caveats() {
         if [[ "$model" == gpt-6-astra ]] && ! jq -e '.metadata.provider_cost_usd | numbers' "$f" >/dev/null 2>&1; then
             astra_estimate=1
         fi
+        case "$model" in
+            deepseek-flash|deepseek-v4-flash|deepseek-v4-flash-vision-exp)
+                if ! jq -e '.metadata.provider_cost_usd | numbers' "$f" >/dev/null 2>&1; then
+                    deepseek_flash_estimate=1
+                fi
+                ;;
+        esac
+        if ! jq -e '.metadata.provider_cost_usd | numbers' "$f" >/dev/null 2>&1; then
+            [[ "$model" != grok-4.7 ]] || grok_47_estimate=1
+            [[ "$model" != claude-opus-5-5 ]] || opus_55_estimate=1
+        fi
+        case "$model" in
+            gemini-3.8-flash|'Gemini 3.8 Flash (High)'|'Gemini 3.8 Flash (Medium)'|'Gemini 3.8 Flash (Low)')
+                if ! jq -e '.metadata.provider_cost_usd | numbers' "$f" >/dev/null 2>&1; then
+                    gemini_38_estimate=1
+                fi
+                ;;
+        esac
         is_unpriced_model "$model" || continue
         seen=false
         for item in "${unpriced_models[@]+"${unpriced_models[@]}"}"; do
@@ -441,6 +471,18 @@ format_cost_caveats() {
     if [[ $astra_estimate -eq 1 ]]; then
         parts="${parts:+$parts; }cost estimated using Astra Standard rates; cache/service adjustments excluded, not an invoice"
     fi
+    if [[ $deepseek_flash_estimate -eq 1 ]]; then
+        parts="${parts:+$parts; }DeepSeek Flash cost estimated at peak cache-miss rates; off-peak/cache discounts excluded, not an invoice"
+    fi
+    if [[ $grok_47_estimate -eq 1 ]]; then
+        parts="${parts:+$parts; }Grok 4.7 API list-rate estimate includes long-context pricing; cache/regional/priority/Fast adjustments excluded, not an invoice"
+    fi
+    if [[ $opus_55_estimate -eq 1 ]]; then
+        parts="${parts:+$parts; }Opus 5.5 Standard API estimate excludes cache/service adjustments, not an invoice"
+    fi
+    if [[ $gemini_38_estimate -eq 1 ]]; then
+        parts="${parts:+$parts; }Gemini 3.8 promotional API estimate through 2026-12-31; cache/service adjustments excluded, not a CLI invoice"
+    fi
     local unpriced=""
     if (( ${#unpriced_models[@]} > 0 )); then
         unpriced=$(IFS=,; printf '%s' "${unpriced_models[*]}")

package/scripts/query_deepseek.sh

@@ -5,7 +5,7 @@
 #
 # Environment variables:
 #   DEEPSEEK_API_KEY   - API key for DeepSeek (required)
-#   DEEPSEEK_MODEL     - Model to use (default: deepseek-reasoner)
+#   DEEPSEEK_MODEL     - Model to use (default: deepseek-flash (DeepSeek-V4.1-Flash))
 #   DEEPSEEK_TIMEOUT   - Timeout in seconds (default: 180)
 #   ENABLE_PERSONA     - Enable "The Code Specialist" persona (default: true)
 

package/scripts/query_gemini.sh

@@ -8,7 +8,7 @@
 # Usage: ./query_gemini.sh "question" [context_file] [output_file]
 #
 # Environment variables:
-#   GEMINI_MODEL - agy model display name (default: "Gemini 3.7 Flash (High)")
+#   GEMINI_MODEL - agy model display name (default: "Gemini 3.8 Flash (High)")
 #   GEMINI_TIMEOUT - Timeout in seconds (default: 180)
 #   GEMINI_USE_API - Use Google AI API mode instead of the agy CLI (default: false)
 #   GEMINI_API_MODEL - API model ID for API mode (default: gemini-3.1-pro-preview)

package/scripts/query_grok.sh

@@ -5,7 +5,7 @@
 #
 # Environment variables:
 #   GROK_CMD        - Grok Build command (default: grok)
-#   GROK_MODEL      - Model to use in both transports (default: grok-4.6)
+#   GROK_MODEL      - Model to use in both transports (default: grok-4.7)
 #   GROK_TIMEOUT    - Timeout in seconds (default: 180)
 #   GROK_MAX_TURNS  - Bounded advisory turns (default: 4)
 #   GROK_USE_API    - Force the API path when true (default: CLI-first auto)