Boarding pass · Best value
MiMo V2.5 Pro
mimo-v2.5-pro · Xiaomi
- Gate
- Agentic coding
- Ratings
- 7
- Recommend
- 100%
- Value · Quality · Speed
- 4.9 · 4.7 · 4.3
Value
4.9
out of 5 for value
CS-2026-01
Now boarding · category leaders
Top of the registry
Highest average in each score across every model with 5+ verified ratings. Ties go to the model with more ratings.
Boarding pass · Best quality
Gemma 4 31B IT
gemma-4-31b-it · Google
- Gate
- Docs & explanations
- Ratings
- 5
- Recommend
- 100%
- Value · Quality · Speed
- 4.4 · 4.8 · 3.8
Quality
4.8
out of 5 for quality
CS-2026-08
Boarding pass · Fastest
Gemini 3.6 Flash
gemini-3.6-flash · Google
- Gate
- Writing tests
- Ratings
- 96
- Recommend
- 89%
- Value · Quality · Speed
- 4.1 · 3.7 · 4.3
Speed
4.3
out of 5 for speed
CS-2026-17
As of October 24, 2026, MiMo V2.5 Pro leads the 2026 DevPass Model Census on value for money (4.9/5 across 7 verified ratings), Gemma 4 31B IT leads on output quality (4.8/5), and Gemini 3.6 Flash leads on speed (4.3/5). 1,124 DevPass developers have filed 1,703 verified ratings across 41 models. The most-rated model is GPT-5.6 Sol with 242 ratings and a 97% recommend rate.
Departures · full registry
Every model on the registry
Click a column to sort, or narrow the board by vendor, use case, and rating count. Filters live in the URL, so a view can be shared.
Showing 41 of 41 models · sorted by value descendingDOC. CS-2026
Departures · Coding models
Rank = registry position by value · refreshed every 5 min
| Registry rank | Status | ||||||
|---|---|---|---|---|---|---|---|
| 1 | MiMo V2.5 Pro Xiaomi · mimo-v2.5-pro · mostly agentic coding | 4.9 out of 5 value | 4.7 out of 5 quality | 4.3 out of 5 speed | 100% | 7 | Cleared: 90% or more of raters would recommend this model |
| 2 | DeepSeek V4.1 Flash DeepSeek · deepseek-v4.1-flash · mostly agentic coding | 4.7 out of 5 value | 4.0 out of 5 quality | 4.3 out of 5 speed | 98% | 65 | Cleared: 90% or more of raters would recommend this model |
| 3 | MiMo V2.6 Flash Xiaomi · mimo-v2.6-flash · mostly agentic coding | 4.6 out of 5 value | 3.6 out of 5 quality | 2.8 out of 5 speed | 80% | 5 | Boarding: 75–89% of raters would recommend this model |
| 4 | MiMo V2.6 Pro Xiaomi · mimo-v2.6-pro · mostly agentic coding | 4.6 out of 5 value | 4.1 out of 5 quality | 3.1 out of 5 speed | 100% | 17 | Cleared: 90% or more of raters would recommend this model |
| 5 | GLM-5.3 Flash Z.ai · glm-5.3-flash · mostly agentic coding | 4.6 out of 5 value | 4.1 out of 5 quality | 3.5 out of 5 speed | 97% | 64 | Cleared: 90% or more of raters would recommend this model |
| 6 | DeepSeek V4 Flash DeepSeek · deepseek-v4-flash · mostly agentic coding | 4.5 out of 5 value | 3.7 out of 5 quality | 4.0 out of 5 speed | 98% | 83 | Cleared: 90% or more of raters would recommend this model |
| 7 | DeepSeek V4 Pro DeepSeek · deepseek-v4-pro · mostly agentic coding | 4.5 out of 5 value | 3.7 out of 5 quality | 3.9 out of 5 speed | 93% | 75 | Cleared: 90% or more of raters would recommend this model |
| 8 | Gemma 4 31B IT Google · gemma-4-31b-it · mostly docs & explanations | 4.4 out of 5 value | 4.8 out of 5 quality | 3.8 out of 5 speed | 100% | 5 | Cleared: 90% or more of raters would recommend this model |
| 9 | Gemini 3.7 Flash Google · gemini-3.7-flash · mostly writing tests | 4.4 out of 5 value | 4.1 out of 5 quality | 4.2 out of 5 speed | 93% | 166 | Cleared: 90% or more of raters would recommend this model |
| 10 | Qwen3.8 Max Alibaba · qwen3.8-max · mostly agentic coding | 4.3 out of 5 value | 4.3 out of 5 quality | 3.5 out of 5 speed | 100% | 15 | Cleared: 90% or more of raters would recommend this model |
| 11 | Gemini 3.8 Flash Google · gemini-3.8-flash · mostly writing tests | 4.3 out of 5 value | 4.0 out of 5 quality | 4.1 out of 5 speed | 94% | 148 | Cleared: 90% or more of raters would recommend this model |
| 12 | GPT-6 Luna OpenAI · gpt-6-luna · mostly agentic coding | 4.3 out of 5 value | 4.1 out of 5 quality | 3.8 out of 5 speed | 82% | 11 | Boarding: 75–89% of raters would recommend this model |
| 13 | Gemini 3 Flash (Preview) Google · gemini-3-flash-preview · mostly other | 4.2 out of 5 value | 3.9 out of 5 quality | 3.9 out of 5 speed | 100% | 38 | Cleared: 90% or more of raters would recommend this model |
| 14 | Muse Spark 1.3 Contributor Meta · muse-spark-1.3-contributor · mostly agentic coding | 4.2 out of 5 value | 3.8 out of 5 quality | 3.3 out of 5 speed | 100% | 10 | Cleared: 90% or more of raters would recommend this model |
| 15 | GLM-5.3 Z.ai · glm-5.3 · mostly agentic coding | 4.2 out of 5 value | 4.2 out of 5 quality | 3.9 out of 5 speed | 100% | 26 | Cleared: 90% or more of raters would recommend this model |
| 16 | GPT-5.6 Luna OpenAI · gpt-5.6-luna · mostly agentic coding | 4.2 out of 5 value | 3.9 out of 5 quality | 4.2 out of 5 speed | 94% | 50 | Cleared: 90% or more of raters would recommend this model |
| 17 | Gemini 3.6 Flash Google · gemini-3.6-flash · mostly writing tests | 4.1 out of 5 value | 3.7 out of 5 quality | 4.3 out of 5 speed | 89% | 96 | Boarding: 75–89% of raters would recommend this model |
| 18 | GPT-5.1 OpenAI · gpt-5.1 · mostly writing tests | 4.1 out of 5 value | 3.9 out of 5 quality | 3.6 out of 5 speed | 88% | 17 | Boarding: 75–89% of raters would recommend this model |
| 19 | GPT-5.6 Sol OpenAI · gpt-5.6-sol · mostly writing tests | 4.0 out of 5 value | 4.3 out of 5 quality | 3.7 out of 5 speed | 97% | 242 | Cleared: 90% or more of raters would recommend this model |
| 20 | Qwen3.8 Flash Alibaba · qwen3.8-flash · mostly agentic coding | 4.0 out of 5 value | 3.8 out of 5 quality | 4.0 out of 5 speed | 90% | 10 | Cleared: 90% or more of raters would recommend this model |
| 21 | Kimi K2.7 Code Moonshot · kimi-k2.7-code · mostly agentic coding | 4.0 out of 5 value | 4.0 out of 5 quality | 3.6 out of 5 speed | 100% | 5 | Cleared: 90% or more of raters would recommend this model |
| 22 | GLM-5.2 Z.ai · glm-5.2 · mostly agentic coding | 4.0 out of 5 value | 4.0 out of 5 quality | 3.6 out of 5 speed | 96% | 71 | Cleared: 90% or more of raters would recommend this model |
| 23 | Gemini 3.5 Flash Lite Google · gemini-3.5-flash-lite · mostly writing tests | 3.9 out of 5 value | 3.3 out of 5 quality | 4.3 out of 5 speed | 69% | 13 | Standby: Fewer than 75% of raters would recommend this model |
| 24 | Claude Fable 5.1 Anthropic · claude-fable-5-1 · mostly agentic coding | 3.9 out of 5 value | 4.4 out of 5 quality | 3.5 out of 5 speed | 88% | 8 | Boarding: 75–89% of raters would recommend this model |
| 25 | Gemini 4 Argon Google · gemini-4-argon · mostly writing tests | 3.9 out of 5 value | 4.0 out of 5 quality | 4.0 out of 5 speed | 75% | 8 | Boarding: 75–89% of raters would recommend this model |
| 26 | GPT-5.6 Terra OpenAI · gpt-5.6-terra · mostly agentic coding | 3.9 out of 5 value | 3.9 out of 5 quality | 4.1 out of 5 speed | 87% | 15 | Boarding: 75–89% of raters would recommend this model |
| 28 | Gemini 3.1 Pro (Preview) Google · gemini-3.1-pro-preview · mostly writing tests | 3.9 out of 5 value | 3.7 out of 5 quality | 3.3 out of 5 speed | 86% | 146 | Boarding: 75–89% of raters would recommend this model |
| 27 | Gemini 3.5 Flash Google · gemini-3.5-flash · mostly writing tests | 3.9 out of 5 value | 3.8 out of 5 quality | 4.0 out of 5 speed | 82% | 22 | Boarding: 75–89% of raters would recommend this model |
| 29 | Claude Opus 5.5 Anthropic · claude-opus-5-5 · mostly agentic coding | 3.8 out of 5 value | 4.3 out of 5 quality | 3.4 out of 5 speed | 100% | 15 | Cleared: 90% or more of raters would recommend this model |
| 30 | GPT-5.5 OpenAI · gpt-5.5 · mostly agentic coding | 3.8 out of 5 value | 4.3 out of 5 quality | 4.0 out of 5 speed | 88% | 8 | Boarding: 75–89% of raters would recommend this model |
| 31 | MiniMax M3 MiniMax · minimax-m3 · mostly agentic coding | 3.7 out of 5 value | 3.7 out of 5 quality | 3.8 out of 5 speed | 92% | 13 | Cleared: 90% or more of raters would recommend this model |
| 32 | Kimi K3 Moonshot · kimi-k3 · mostly agentic coding | 3.6 out of 5 value | 4.4 out of 5 quality | 3.9 out of 5 speed | 98% | 46 | Cleared: 90% or more of raters would recommend this model |
| 34 | Claude Opus 4.6 Anthropic · claude-opus-4-6 · mostly writing tests | 3.6 out of 5 value | 4.1 out of 5 quality | 3.4 out of 5 speed | 96% | 72 | Cleared: 90% or more of raters would recommend this model |
| 33 | Claude Sonnet 5 Anthropic · claude-sonnet-5 · mostly agentic coding | 3.6 out of 5 value | 4.0 out of 5 quality | 3.6 out of 5 speed | 88% | 25 | Boarding: 75–89% of raters would recommend this model |
| 35 | GPT-6 Astra OpenAI · gpt-6-astra · mostly autocomplete | 3.5 out of 5 value | 4.5 out of 5 quality | 4.2 out of 5 speed | 100% | 6 | Cleared: 90% or more of raters would recommend this model |
| 36 | Gemini 3.1 Flash Lite Google · gemini-3.1-flash-lite · mostly writing tests | 3.4 out of 5 value | 3.4 out of 5 quality | 4.1 out of 5 speed | 90% | 10 | Cleared: 90% or more of raters would recommend this model |
| 37 | Grok 4.5 xAI · grok-4-5 · mostly agentic coding | 3.3 out of 5 value | 3.7 out of 5 quality | 4.0 out of 5 speed | 92% | 12 | Cleared: 90% or more of raters would recommend this model |
| 39 | Claude Opus 5 Anthropic · claude-opus-5 · mostly agentic coding | 3.3 out of 5 value | 4.3 out of 5 quality | 3.3 out of 5 speed | 91% | 23 | Cleared: 90% or more of raters would recommend this model |
| 38 | Claude Fable 5 Anthropic · claude-fable-5 · mostly agentic coding | 3.3 out of 5 value | 4.4 out of 5 quality | 3.3 out of 5 speed | 89% | 19 | Boarding: 75–89% of raters would recommend this model |
| 40 | Claude Opus 4.8 Anthropic · claude-opus-4-8 · mostly agentic coding | 3.0 out of 5 value | 4.2 out of 5 quality | 3.8 out of 5 speed | 91% | 11 | Cleared: 90% or more of raters would recommend this model |
| 41 | Claude Sonnet 4.6 Anthropic · claude-sonnet-4-6 · mostly agentic coding | 3.0 out of 5 value | 4.2 out of 5 quality | 3.4 out of 5 speed | 80% | 5 | Boarding: 75–89% of raters would recommend this model |
- Value
- 4.4
- Quality
- 4.1
- Speed
- 4.2
Mostly writing testsCleared: 90% or more of raters would recommend this model- Value
- 4.3
- Quality
- 4.3
- Speed
- 3.5
Mostly agentic codingCleared: 90% or more of raters would recommend this model- Value
- 4.3
- Quality
- 4.0
- Speed
- 4.1
Mostly writing testsCleared: 90% or more of raters would recommend this model- Value
- 4.3
- Quality
- 4.1
- Speed
- 3.8
Mostly agentic codingBoarding: 75–89% of raters would recommend this model- Value
- 4.2
- Quality
- 3.9
- Speed
- 3.9
Mostly otherCleared: 90% or more of raters would recommend this model- Value
- 4.2
- Quality
- 3.8
- Speed
- 3.3
Mostly agentic codingCleared: 90% or more of raters would recommend this model- Value
- 4.2
- Quality
- 4.2
- Speed
- 3.9
Mostly agentic codingCleared: 90% or more of raters would recommend this model- Value
- 4.2
- Quality
- 3.9
- Speed
- 4.2
Mostly agentic codingCleared: 90% or more of raters would recommend this model- Value
- 4.1
- Quality
- 3.7
- Speed
- 4.3
Mostly writing testsBoarding: 75–89% of raters would recommend this model- Value
- 4.1
- Quality
- 3.9
- Speed
- 3.6
Mostly writing testsBoarding: 75–89% of raters would recommend this model- Value
- 4.0
- Quality
- 4.3
- Speed
- 3.7
Mostly writing testsCleared: 90% or more of raters would recommend this model- Value
- 4.0
- Quality
- 3.8
- Speed
- 4.0
Mostly agentic codingCleared: 90% or more of raters would recommend this model- Value
- 4.0
- Quality
- 4.0
- Speed
- 3.6
Mostly agentic codingCleared: 90% or more of raters would recommend this model- Value
- 4.0
- Quality
- 4.0
- Speed
- 3.6
Mostly agentic codingCleared: 90% or more of raters would recommend this model- Value
- 3.9
- Quality
- 3.3
- Speed
- 4.3
Mostly writing testsStandby: Fewer than 75% of raters would recommend this model- Value
- 3.9
- Quality
- 4.4
- Speed
- 3.5
Mostly agentic codingBoarding: 75–89% of raters would recommend this model- Value
- 3.9
- Quality
- 4.0
- Speed
- 4.0
Mostly writing testsBoarding: 75–89% of raters would recommend this model- Value
- 3.9
- Quality
- 3.9
- Speed
- 4.1
Mostly agentic codingBoarding: 75–89% of raters would recommend this model- Value
- 3.9
- Quality
- 3.7
- Speed
- 3.3
Mostly writing testsBoarding: 75–89% of raters would recommend this model- Value
- 3.9
- Quality
- 3.8
- Speed
- 4.0
Mostly writing testsBoarding: 75–89% of raters would recommend this model- Value
- 3.8
- Quality
- 4.3
- Speed
- 3.4
Mostly agentic codingCleared: 90% or more of raters would recommend this model- Value
- 3.8
- Quality
- 4.3
- Speed
- 4.0
Mostly agentic codingBoarding: 75–89% of raters would recommend this model- Value
- 3.7
- Quality
- 3.7
- Speed
- 3.8
Mostly agentic codingCleared: 90% or more of raters would recommend this model- Value
- 3.6
- Quality
- 4.4
- Speed
- 3.9
Mostly agentic codingCleared: 90% or more of raters would recommend this model- Value
- 3.6
- Quality
- 4.1
- Speed
- 3.4
Mostly writing testsCleared: 90% or more of raters would recommend this model- Value
- 3.6
- Quality
- 4.0
- Speed
- 3.6
Mostly agentic codingBoarding: 75–89% of raters would recommend this model- Value
- 3.5
- Quality
- 4.5
- Speed
- 4.2
Mostly autocompleteCleared: 90% or more of raters would recommend this model- Value
- 3.4
- Quality
- 3.4
- Speed
- 4.1
Mostly writing testsCleared: 90% or more of raters would recommend this model- Value
- 3.3
- Quality
- 3.7
- Speed
- 4.0
Mostly agentic codingCleared: 90% or more of raters would recommend this model- Value
- 3.3
- Quality
- 4.3
- Speed
- 3.3
Mostly agentic codingCleared: 90% or more of raters would recommend this model- Value
- 3.3
- Quality
- 4.4
- Speed
- 3.3
Mostly agentic codingBoarding: 75–89% of raters would recommend this model- Value
- 3.0
- Quality
- 4.2
- Speed
- 3.8
Mostly agentic codingCleared: 90% or more of raters would recommend this model- Value
- 3.0
- Quality
- 4.2
- Speed
- 3.4
Mostly agentic codingBoarding: 75–89% of raters would recommend this model
Scores are 1–5 averages. Models need 5+ verified ratings to be listed; totals refresh every few minutes. Last updated October 24, 2026 (UTC).
Conditions of entry
The rules of the registry
Usage-verified, or it doesn't count
You can only rate a model your DevPass workspace has hit with 50+ requests in the past 30 days. Nobody rates a model they read a thread about.
Members only
Every respondent has an active, paid DevPass plan. These are verdicts from people spending their own credits.
No small-sample noise
A model is published only after 5 or more developers rate it, and only aggregates ever leave the building.
One reward per member per quarter
The census runs in quarterly waves. Your first entry of each wave earns a free Reset Pass — rate as many models as you use, but nobody can farm passes.
Reading the census
Model Census FAQ
- How are models scored in the DevPass Model Census?
- Each entry rates one model from 1 to 5 on value for money, output quality, and speed, plus a yes-or-no “would you recommend it”. The registry shows the per-model averages and the share of raters who would recommend the model.
- Who can rate a model?
- Only paid DevPass members, and only for models their workspace has sent at least 50 requests to in the past 30 days. Ratings are tied to verified usage, not opinions from a thread.
- How is the registry ranked?
- Registry rank is the model's position by average value score, with ties broken by number of ratings. The sort and filter controls change what you see on the board but never the underlying rank.
- Why is a model missing from the registry?
- A model appears once 5 or more developers have filed a verified rating on it in 2026. Per-use-case breakdowns are additionally hidden below two entries so no single respondent can be identified.
- How often does the census update?
- Members file entries in quarterly waves and the yearly registry aggregates every wave. The published totals refresh every five minutes.
Your entryWave Q4 · 2026
Rate the models you ship with, stamp a free Reset Pass
Takes two minutes. Your first entry of the wave earns a Reset Pass on your tier, and every entry sharpens the registry for the next developer at the desk.