MiniMax M2.7 vs Sakana Namazu: which should you choose in 2026?
MiniMax M2.7 (by MiniMax) and Sakana Namazu (by Sakana AI) are compared below. Here is how they stack up on benchmarks, price, and capabilities, and which one to pick in 2026.
There are not enough shared, protocol-compatible benchmark results to declare a performance leader.
MiniMax M2.7 is about 3.3× cheaper on a blended 3:1 input/output basis ($0.525 vs $1.71 per 1M tokens).
Sakana Namazu has the larger context window (262,144 tokens vs 204,800 tokens).
Choose MiniMax M2.7 if…
- • lower blended API cost matters for your workload
Choose Sakana Namazu if…
- • you work with longer documents, transcripts, or codebases
- • you need image inputs
The benchmark count includes only results measured with a matching benchmark version and protocol. Arena scores are shown separately. Missing, preliminary, and incompatible data is not treated as a controlled win. Published point-score comparisons are labeled separately when protocol details are incomplete. For text-output models, the verdict also compares token pricing and context windows.
Performance benchmarks
Every value links to its source. A dash means that no reviewed result is available for that exact model and protocol.
| Benchmark | MiniMax M2.7 | Sakana Namazu |
|---|---|---|
A 657-task benchmark of multi-step work across simulated SaaS applications in six business domains. The Artificial Analysis protocol reports a guardrail-aware score and is distinct from both the public Zapier split and the unrelated dynamic AutoBench framework. Guardrail-aware score | — | — |
A benchmark of difficult, verifiable information-seeking questions designed to measure an agent's ability to locate hard-to-find facts through web browsing. Accuracy | — | — |
A machine-learning engineering benchmark. MiniMax reports the average medal rate across three trials for M2.7. Average medal rate | 66.6% MiniMax report | — |
A multimodal agent benchmark reported by MiniMax, measuring accuracy under the provider's disclosed launch evaluation setup. Accuracy | 62.7% MiniMax report | — |
A verified computer-use benchmark in which multimodal agents operate desktop applications and are graded from the resulting environment state. Mean task reward | — | — |
A tool-use benchmark reported by MiniMax. The provider result is source-attributed but not treated as a controlled comparison without a matched protocol. Success rate | 46.3% MiniMax report | — |
A multi-language and multi-repository software-engineering benchmark reported by MiniMax, retained with its source-specific protocol identity. Resolved | 52.7% MiniMax report | — |
A repository-generation benchmark reported in the MiniMax M2.7 launch evaluation, with incomplete public harness disclosure in that source. Score | 39.8% MiniMax report | — |
A multilingual software-engineering benchmark reported by MiniMax. The launch source does not disclose a complete evaluation protocol. Resolved | 76.5% MiniMax report | — |
A contamination-resistant software-engineering benchmark with long-horizon tasks across multiple programming languages. Public and private splits are distinct protocols. Resolved | 56.2% MiniMax report | — |
A human-validated subset of real GitHub issues used to measure whether a coding agent can produce repository patches that resolve the associated tests. Resolved | — | — |
The Terminal-Bench 2 result named in MiniMax's launch report. Its exact minor version and harness are not disclosed, so it must remain separate from versioned Terminal-Bench 2.0 and 2.1 rows. Pass rate | 57.0% MiniMax report | — |
Version 2.1 of the benchmark for completing realistic tasks in terminal environments. Harness, resource limits, and attempt count are part of the protocol. Mean task success | — | — |
A provider-reported benchmark of long-horizon software-engineering work. The MiniMax launch source does not disclose enough harness detail for controlled cross-provider comparison. Score | 55.6% MiniMax report | — |
A dynamic LLM evaluation framework in which models generate questions, answer them, and participate in reciprocal peer assessment. AutoBench is distinct from Zapier's AutomationBench. Weighted peer-assessment score | — | — |
An Artificial Analysis GDPval evaluation reported as an Elo rating in MiniMax's official M2.7 launch post. Elo | 1,495 MiniMax report | — |
The highest-quality subset of Graduate-Level Google-Proof Q&A, designed to test expert-level scientific reasoning in biology, physics, and chemistry. Accuracy | — | — |
A 2,500-question expert-level benchmark spanning dozens of academic fields. Tool-assisted and no-tools results are separate protocols and must not be merged. Accuracy | — | — |
| Arena preference scores | ||
Human preference score | 1416 ±4 | — |
Human preference score for code and web development | 1398 ±6 | — |
Benchmark sources
Reviewed evidence last updated 2026-08-09. Arena data last refreshed Aug 25, 2026.
Pricing, capabilities, and model facts
| Feature | MiniMax M2.7 | Sakana Namazu |
|---|---|---|
| Context & model facts | ||
| Developer | MiniMax | Sakana AI |
| API provider | MiniMax | Sakana |
| Input context | 204,800 tokens | 262,144 tokens |
| Maximum output | 131,072 tokens | 65,536 tokens |
| Released | Mar 18, 2026 | — |
| Added to Writingmate | Mar 18, 2026 | Aug 11, 2026 |
| License | Not available | Not available |
| Knowledge cutoff | — | — |
| Capabilities | ||
| Inputs | Text | Text, Image, File |
| Outputs | Text | Text |
| Provider endpoint accepts tool parameters | Yes | Yes |
| Reasoning | Yes | Yes |
| Vision | No | Yes |
| Image Generation | No | No |
| Video Generation | No | No |
| API pricing | ||
| Input (per 1M tokens) | $0.30 | $0.95 |
| Output (per 1M tokens) | $1.20 | $4.00 |
| Blended 3:1 input/output | $0.525 | $1.71 |
| API performance | ||
| p95 latency | Not measured | Not measured |
| Output throughput | Not measured | Not measured |
| Writingmate shows API performance only when both models have enough observations from the same measurement window, prompt profile, and provider. Third-party latency values are not copied into this table. | ||
Data sources
- MiniMax M2.7: Early Echoes of Self-Evolution
- Writingmate model catalog (pricing, limits, and availability)
Catalog data last updated Aug 11, 2026.
Why Pay for Multiple Subscriptions?
Comparing MiniMax M2.7 from MiniMax with Sakana Namazu from Sakana AI? Instead of managing separate API keys and subscriptions, get both with Writingmate.
| Plan | Price | MiniMax M2.7 | Sakana Namazu | AI Images | AI Video |
|---|---|---|---|---|---|
Writingmate Pro Most popular | $20/mo | Included | Included | Nano Banana Pro, FLUX.2, DALL-E & more | Sora 2, VEO 3.1 |
Writingmate Ultimate Power users | $60/mo | Included | Included | Nano Banana Pro, FLUX.2, DALL-E & more | Sora 2, VEO 3.1 |
MiniMax M2.7 vs Sakana Namazu FAQ
Which is better, MiniMax M2.7 or Sakana Namazu?
There are not enough shared, protocol-compatible benchmark results to declare a performance leader. MiniMax M2.7 is about 3.3× cheaper on a blended 3:1 input/output basis ($0.525 vs $1.71 per 1M tokens). Sakana Namazu has the larger context window (262,144 tokens vs 204,800 tokens).
Which model is cheaper to use through an API?
MiniMax M2.7 is about 3.3× cheaper on a blended 3:1 input/output basis ($0.525 vs $1.71 per 1M tokens).
Which model supports more context?
Sakana Namazu has the larger context window (262,144 tokens vs 204,800 tokens).
Can I switch between MiniMax M2.7 and Sakana Namazu?
Yes. Use the model selector on this page to open any current Writingmate model comparison. You can also run the same prompt with both models in Writingmate.
Related comparisons
More MiniMax M2.7 comparisons
Muse Spark 1.2 Contributor
Meta
Muse Spark 1.2 contributor tier is a reasoning model from Meta designed for developers who want to start building at an even lower cost.
DeepSeek V4 Flash Vision Exp
DeepSeek
DeepSeek V4 Flash Vision Exp is a model by DeepSeek.
Ox Alpha
Stealth
Ox Alpha is a reasoning model designed for coding, sustained agentic work, and production workloads.
GLM 5.3
Z.AI
GLM-5.3 is a large-scale reasoning model from Z.ai, built for complex software engineering and long-horizon agent tasks.
Qwen3.8 27B
Qwen
Qwen3.8 27B is an open-weight dense vision-language model from Qwen.
Gemini 3.7 Flash
Gemini 3.7 Flash is a multimodal model from Google for fast agentic workflows, coding, and complex multi-step reasoning.
More Sakana Namazu comparisons
Muse Spark 1.2 Contributor
Meta
Muse Spark 1.2 contributor tier is a reasoning model from Meta designed for developers who want to start building at an even lower cost.
Gemini 3.7 Flash
Gemini 3.7 Flash is a multimodal model from Google for fast agentic workflows, coding, and complex multi-step reasoning.
Grok 4.6
xAI
Grok 4.6 is SpaceXAI's smartest model with frontier performance on coding, knowledge work, and STEM.
Muse Spark 1.2
Meta
Muse Spark 1.2 is a reasoning model from Meta, designed for complex agentic tasks.
Claude Opus 5
Anthropic
Claude Opus 5 is Anthropic’s flagship model for demanding reasoning, coding, and long-horizon agentic work.
Gemini 3.6 Flash
Gemini 3.6 Flash is a high-efficiency model from Google for coding, agentic workflows, and web and app development.