Language modelActive
Fugu Max
Sakana AI
- Released
- September 11, 2026
- Data date
- October 7, 2026
- Price
- Input · Output / 1M tokens
- Input
- Text, image
- Output
- Text
Fugu Max expands Sakana Fugu with a larger pool of open and specialist models. The system dynamically routes tasks to an efficient model instead of handling every request with one fixed model.
Sakana offers Fugu Max through an OpenAI-compatible API. The fugu-max alias currently points to fugu-max-v1.0; input and output cost $2 and $6 per 1M tokens, with cached input at $0.25.
Specifications and access
| Specification | Value and source |
|---|---|
| API model ID | fugu-maxSource |
| Pinned model version | fugu-max-v1.0; current target of the fugu-max aliasSource |
| System type | Multi-agent orchestration that dynamically routes tasks to suitable models in its agent poolSource |
| Input modalities | Text, imageSource |
| Output modalities | TextSource |
| Availability | Available through the Sakana Fugu API since September 11, 2026Source |
Page 1 of 3
Measurements without matching peer values
These measurements have no matching peer values under the same test conditions. Their original values and sources remain available here.
Agronomy LLM Benchmark: Overall Score (Show measurement, test conditions, and source)
- Source value
- 87.53
- Score
- 87.53
- Metric
- Evaluator-reported accuracy
- Unit
- %
- Category
- Agronomy LLM Benchmark / Overall Score
- Direction
- Higher is better
- Source type
- independent-evaluator
- Evaluator
- gbstox/agronomy_llm_benchmarking README leaderboard
- Status
- active
- Retrieved at
- 2026-10-07
- Methodology
- Independent evaluation. OpenRouter route/model ID=sakana/fugu-max; evaluator's tested date=2026-09-23; benchmark=agronomy multiple-choice; request settings and immutable backend version not stated. The evaluator table identifies the model as proprietary.
- Context
- This is an independent evaluator's exact OpenRouter route result on its agronomy multiple-choice benchmark, not a general-purpose capability score. The evaluator keeps questions, images and per-model raw outputs private. The README says the image-domain columns are excluded from the overall score and does not publish Fugu's immutable backend version or request settings.
Agronomy LLM Benchmark: Community Questions FBN (Show measurement, test conditions, and source)
- Source value
- 80
- Score
- 80
- Metric
- Evaluator-reported accuracy
- Unit
- %
- Category
- Agronomy LLM Benchmark / Community Questions FBN
- Direction
- Higher is better
- Source type
- independent-evaluator
- Evaluator
- gbstox/agronomy_llm_benchmarking README leaderboard
- Status
- active
- Retrieved at
- 2026-10-07
- Methodology
- Independent evaluation. OpenRouter route/model ID=sakana/fugu-max; evaluator's tested date=2026-09-23; benchmark=agronomy multiple-choice; request settings and immutable backend version not stated. The evaluator table identifies the model as proprietary.
- Context
- This is an independent evaluator's exact OpenRouter route result on its agronomy multiple-choice benchmark, not a general-purpose capability score. The evaluator keeps questions, images and per-model raw outputs private. The README says the image-domain columns are excluded from the overall score and does not publish Fugu's immutable backend version or request settings.
Agronomy LLM Benchmark: Crop Management (Show measurement, test conditions, and source)
- Source value
- 88.2
- Score
- 88.2
- Metric
- Evaluator-reported accuracy
- Unit
- %
- Category
- Agronomy LLM Benchmark / Crop Management
- Direction
- Higher is better
- Source type
- independent-evaluator
- Evaluator
- gbstox/agronomy_llm_benchmarking README leaderboard
- Status
- active
- Retrieved at
- 2026-10-07
- Methodology
- Independent evaluation. OpenRouter route/model ID=sakana/fugu-max; evaluator's tested date=2026-09-23; benchmark=agronomy multiple-choice; request settings and immutable backend version not stated. The evaluator table identifies the model as proprietary.
- Context
- This is an independent evaluator's exact OpenRouter route result on its agronomy multiple-choice benchmark, not a general-purpose capability score. The evaluator keeps questions, images and per-model raw outputs private. The README says the image-domain columns are excluded from the overall score and does not publish Fugu's immutable backend version or request settings.
Agronomy LLM Benchmark: Image Disease (Show measurement, test conditions, and source)
- Source value
- 56
- Score
- 56
- Metric
- Evaluator-reported accuracy
- Unit
- %
- Category
- Agronomy LLM Benchmark / Image Disease
- Direction
- Higher is better
- Source type
- independent-evaluator
- Evaluator
- gbstox/agronomy_llm_benchmarking README leaderboard
- Status
- active
- Retrieved at
- 2026-10-07
- Methodology
- Independent evaluation. OpenRouter route/model ID=sakana/fugu-max; evaluator's tested date=2026-09-23; benchmark=agronomy multiple-choice; request settings and immutable backend version not stated. The evaluator table identifies the model as proprietary.
- Context
- This is an independent evaluator's exact OpenRouter route result on its agronomy multiple-choice benchmark, not a general-purpose capability score. The evaluator keeps questions, images and per-model raw outputs private. The README says the image-domain columns are excluded from the overall score and does not publish Fugu's immutable backend version or request settings.
Agronomy LLM Benchmark: Image Pest (Show measurement, test conditions, and source)
- Source value
- 58
- Score
- 58
- Metric
- Evaluator-reported accuracy
- Unit
- %
- Category
- Agronomy LLM Benchmark / Image Pest
- Direction
- Higher is better
- Source type
- independent-evaluator
- Evaluator
- gbstox/agronomy_llm_benchmarking README leaderboard
- Status
- active
- Retrieved at
- 2026-10-07
- Methodology
- Independent evaluation. OpenRouter route/model ID=sakana/fugu-max; evaluator's tested date=2026-09-23; benchmark=agronomy multiple-choice; request settings and immutable backend version not stated. The evaluator table identifies the model as proprietary.
- Context
- This is an independent evaluator's exact OpenRouter route result on its agronomy multiple-choice benchmark, not a general-purpose capability score. The evaluator keeps questions, images and per-model raw outputs private. The README says the image-domain columns are excluded from the overall score and does not publish Fugu's immutable backend version or request settings.
Agronomy LLM Benchmark: Image Weed (Show measurement, test conditions, and source)
- Source value
- 72
- Score
- 72
- Metric
- Evaluator-reported accuracy
- Unit
- %
- Category
- Agronomy LLM Benchmark / Image Weed
- Direction
- Higher is better
- Source type
- independent-evaluator
- Evaluator
- gbstox/agronomy_llm_benchmarking README leaderboard
- Status
- active
- Retrieved at
- 2026-10-07
- Methodology
- Independent evaluation. OpenRouter route/model ID=sakana/fugu-max; evaluator's tested date=2026-09-23; benchmark=agronomy multiple-choice; request settings and immutable backend version not stated. The evaluator table identifies the model as proprietary.
- Context
- This is an independent evaluator's exact OpenRouter route result on its agronomy multiple-choice benchmark, not a general-purpose capability score. The evaluator keeps questions, images and per-model raw outputs private. The README says the image-domain columns are excluded from the overall score and does not publish Fugu's immutable backend version or request settings.
Agronomy LLM Benchmark: Nutrient Management (Show measurement, test conditions, and source)
- Source value
- 88.2
- Score
- 88.2
- Metric
- Evaluator-reported accuracy
- Unit
- %
- Category
- Agronomy LLM Benchmark / Nutrient Management
- Direction
- Higher is better
- Source type
- independent-evaluator
- Evaluator
- gbstox/agronomy_llm_benchmarking README leaderboard
- Status
- active
- Retrieved at
- 2026-10-07
- Methodology
- Independent evaluation. OpenRouter route/model ID=sakana/fugu-max; evaluator's tested date=2026-09-23; benchmark=agronomy multiple-choice; request settings and immutable backend version not stated. The evaluator table identifies the model as proprietary.
- Context
- This is an independent evaluator's exact OpenRouter route result on its agronomy multiple-choice benchmark, not a general-purpose capability score. The evaluator keeps questions, images and per-model raw outputs private. The README says the image-domain columns are excluded from the overall score and does not publish Fugu's immutable backend version or request settings.
Agronomy LLM Benchmark: Pest Management (Show measurement, test conditions, and source)
- Source value
- 85.5
- Score
- 85.5
- Metric
- Evaluator-reported accuracy
- Unit
- %
- Category
- Agronomy LLM Benchmark / Pest Management
- Direction
- Higher is better
- Source type
- independent-evaluator
- Evaluator
- gbstox/agronomy_llm_benchmarking README leaderboard
- Status
- active
- Retrieved at
- 2026-10-07
- Methodology
- Independent evaluation. OpenRouter route/model ID=sakana/fugu-max; evaluator's tested date=2026-09-23; benchmark=agronomy multiple-choice; request settings and immutable backend version not stated. The evaluator table identifies the model as proprietary.
- Context
- This is an independent evaluator's exact OpenRouter route result on its agronomy multiple-choice benchmark, not a general-purpose capability score. The evaluator keeps questions, images and per-model raw outputs private. The README says the image-domain columns are excluded from the overall score and does not publish Fugu's immutable backend version or request settings.
Agronomy LLM Benchmark: Soil And Water (Show measurement, test conditions, and source)
- Source value
- 91.2
- Score
- 91.2
- Metric
- Evaluator-reported accuracy
- Unit
- %
- Category
- Agronomy LLM Benchmark / Soil And Water
- Direction
- Higher is better
- Source type
- independent-evaluator
- Evaluator
- gbstox/agronomy_llm_benchmarking README leaderboard
- Status
- active
- Retrieved at
- 2026-10-07
- Methodology
- Independent evaluation. OpenRouter route/model ID=sakana/fugu-max; evaluator's tested date=2026-09-23; benchmark=agronomy multiple-choice; request settings and immutable backend version not stated. The evaluator table identifies the model as proprietary.
- Context
- This is an independent evaluator's exact OpenRouter route result on its agronomy multiple-choice benchmark, not a general-purpose capability score. The evaluator keeps questions, images and per-model raw outputs private. The README says the image-domain columns are excluded from the overall score and does not publish Fugu's immutable backend version or request settings.
Agronomy LLM Benchmark: V1 Benchmark Questions (Show measurement, test conditions, and source)
- Source value
- 86.7
- Score
- 86.7
- Metric
- Evaluator-reported accuracy
- Unit
- %
- Category
- Agronomy LLM Benchmark / V1 Benchmark Questions
- Direction
- Higher is better
- Source type
- independent-evaluator
- Evaluator
- gbstox/agronomy_llm_benchmarking README leaderboard
- Status
- active
- Retrieved at
- 2026-10-07
- Methodology
- Independent evaluation. OpenRouter route/model ID=sakana/fugu-max; evaluator's tested date=2026-09-23; benchmark=agronomy multiple-choice; request settings and immutable backend version not stated. The evaluator table identifies the model as proprietary.
- Context
- This is an independent evaluator's exact OpenRouter route result on its agronomy multiple-choice benchmark, not a general-purpose capability score. The evaluator keeps questions, images and per-model raw outputs private. The README says the image-domain columns are excluded from the overall score and does not publish Fugu's immutable backend version or request settings.
Sakana Fugu Max v1.0: Terminal Bench 2.1 (Show measurement, test conditions, and source)
- Source value
- 89.5
- Score
- 89.5
- Metric
- Vendor-reported orchestration-system benchmark score; chart does not specify a unit
- Unit
- No unit provided
- Category
- Sakana AI Fugu Max v1.0 vendor chart
- Direction
- Higher is better
- Source type
- vendor-reported
- Evaluator
- Sakana AI, Fugu Max v1.0 benchmark chart
- Status
- active
- Retrieved at
- 2026-10-07
- Methodology
- Vendor-reported result. System=Fugu Max v1.0; this measures the vendor's coordinated orchestration system; the chart does not publish a unit, exact harness configuration, or isolated base-model result.
- Context
- This is a vendor-reported result for the end-to-end Fugu orchestration system, not a score for an isolated underlying checkpoint. Sakana's chart does not state the score unit, exact benchmark protocol, model-pool composition for this run, or per-run uncertainty. Treat the value as the vendor's displayed system score only.
Sakana Fugu Max v1.0: HLE (text) (Show measurement, test conditions, and source)
- Source value
- 44.7
- Score
- 44.7
- Metric
- Vendor-reported orchestration-system benchmark score; chart does not specify a unit
- Unit
- No unit provided
- Category
- Sakana AI Fugu Max v1.0 vendor chart
- Direction
- Higher is better
- Source type
- vendor-reported
- Evaluator
- Sakana AI, Fugu Max v1.0 benchmark chart
- Status
- active
- Retrieved at
- 2026-10-07
- Methodology
- Vendor-reported result. System=Fugu Max v1.0; this measures the vendor's coordinated orchestration system; the chart does not publish a unit, exact harness configuration, or isolated base-model result.
- Context
- This is a vendor-reported result for the end-to-end Fugu orchestration system, not a score for an isolated underlying checkpoint. Sakana's chart does not state the score unit, exact benchmark protocol, model-pool composition for this run, or per-run uncertainty. Treat the value as the vendor's displayed system score only.
Page 1 of 3
Sources and data date
| Type | Evidence and data date |
|---|---|
| Price | Base API prices without caching or batch discounts. Higher context tiers and other rates are listed under Costs. |
| Input | Documented modalities. Retrieved October 7, 2026. |
| Output | Documented modalities. Retrieved October 7, 2026. |
| Research status | Research date October 7, 2026. 9 source URLs checked. This documents the inspected sources, not an exhaustive inventory of every publication. |
| Not found in the inspected sources | reproduction of Sakana release-chart protocols Original wording OpenRouter independently evaluated the exact public route sakana/fugu-max on GPQA Diamond, MMMU-Pro Vision, Tau³-Bench Banking, SWE-Atlas QA and DeepSWE. These are OpenRouter's own routed-system runs with separately reported accuracy and epochs, not reproductions of Sakana's release-chart harness. The checked Agronomy evaluation is another independent, task-specific run. Sakana's private model pool and the release chart's full harness remain undisclosed. |
| Not published | immutable backend version for OpenRouter routes Original wording The Agronomy evaluator and OpenRouter's public runs report model_permaslug=sakana/fugu-max rather than an immutable fugu-max-v1.0 pin. Sakana's catalog maps fugu-max to fugu-max-v1.0 at the checked catalog state, but neither evaluator records a provider-side immutable backend revision. |
| Additional source | Sakana Fugu Max release (retrieved October 7, 2026) · Sakana Fugu Max release · Editorial description reviewed October 7, 2026 |
| Additional source | Sakana AI, Fugu Max v1.0 benchmark chart (retrieved October 7, 2026) |
| Additional source | gbstox/agronomy_llm_benchmarking README leaderboard (retrieved October 7, 2026) |
| Additional source | OpenRouter Model Router Benchmarks (retrieved October 7, 2026) |
| Additional source | Model metadata (retrieved October 7, 2026) |
| Additional source | Model metadata (retrieved October 7, 2026) |
| Additional source | Model metadata (retrieved October 7, 2026) |