AI comparison · November 2026
AI model leaderboard — November 2026
58 models from 32 providers. Ratings come from blind votes, prices from each provider's own page, and speed and uptime from real traffic. Measured first-token time, tokens per second and success come from OpenVibe.Report.
OpenVibe.Bench is unavailable. Showing the last good copy, read 2026-11-13 06:35 UTC.
- models listed
- 58
- providers
- 32
- measured endpoints
- 6
Top picks
- Fastestgpt-5-nano1.6 s typical replyFastest typical reply on real traffic this week.
- Best freedeepseek-flash$1.20 per 1M outFree to use now through the Exchange.
Not enough data yet for Best overall, Best value, Best open weights, Best for code, Best for writing, Best uncensored 18+.
Harness and model, apart
A coding agent is a harness driving a model, and the harness can move a score as much as the model. Bench fits the two apart, with an interval on each harness effect. Every agent number here names its harness.
Leaderboard55
No model has blind votes or runs yet, so nothing is ranked: these are the models with their prices and measured speed. Vote in the Arena to rate them.
| Model | Price | First token p50 | Tokens/s p50 | Success | Reply time |
|---|---|---|---|---|---|
| gpt-5-nanoOpenAI Routed here | $0.40 out · $0.05 inper 1M tokens, via OpenAI | — | — | — | 1.6 stypical reply · p95 11 s |
| muse-ember-1.0Meta Routed here | $0.50 out · $0.10 inper 1M tokens, via Meta Model API | — | — | — | No traffic yet |
| gpt-4o-miniOpenAI Routed here | $0.60 out · $0.15 inper 1M tokens, via OpenAI | — | — | — | No traffic yet |
| deepseek-flashDeepSeek Routed here | $1.20 out · $0.30 inper 1M tokens, via DeepSeek | — | — | — | 13 stypical reply · p95 26 s |
| gpt-4.1-miniOpenAI Routed here | $1.60 out · $0.40 inper 1M tokens, via OpenAI | — | — | — | No traffic yet |
| gpt-5-miniOpenAI Routed here | $2.00 out · $0.25 inper 1M tokens, via OpenAI | — | — | — | 2.7 stypical reply · p95 11 s |
| deepseek-v4-proDeepSeek Routed here | $3.96 out · $1.32 inper 1M tokens, via DeepSeek | — | — | — | No traffic yet |
| muse-spark-1.3Meta Routed here | $4.25 out · $1.25 inper 1M tokens, via Meta Model API | — | — | — | No traffic yet |
| gpt-5OpenAI Routed here | $10.00 out · $1.25 inper 1M tokens, via OpenAI | — | — | — | No traffic yet |
| autoopenrouter Not routed here | Free out · Free inper 1M tokens, via IteraCompute | — | — | — | No traffic yet |
| glm-4.7-flashZhipu AI Not routed hereOpen weights | Free out · Free inper 1M tokens, cheapest of 2 providers | — | — | — | No traffic yet |
| meta-llama-3.1-8b-instructMeta Not routed hereOpen weights | $0.08 out · $0.05 inper 1M tokens, via Featherless AI | — | — | — | No traffic yet |
| meta-llama-3.1-8b-instruct-abliteratedMeta Not routed here18+ | $0.08 out · $0.05 inper 1M tokens, via Featherless AI | — | — | — | No traffic yet |
| qwen3-0.6bQwen Not routed here | $0.08 out · $0.04 inper 1M tokens, via Featherless AI | — | — | — | No traffic yet |
| mistral-nemo-instruct-2407-teeMistral AI Not routed here | $0.098 out · $0.025 inper 1M tokens, via Chutes | — | — | — | No traffic yet |
| qwen3.7-flash-2026-07-15Qwen Not routed here | $0.13 out · $0.03 inper 1M tokens, via Qwen (Alibaba Model Studio) | — | — | — | No traffic yet |
| gpt-oss-120bOpenAI Not routed hereOpen weights | $0.17 out · $0.037 inper 1M tokens, cheapest of 7 providers | — | — | — | No traffic yet |
| deepseek-v4-flashDeepSeek Not routed hereOpen weights | $0.18 out · $0.09 inper 1M tokens, cheapest of 2 providers | — | — | — | No traffic yet |
| qwen3-14bQwen Not routed hereOpen weights | $0.24 out · $0.12 inper 1M tokens, cheapest of 2 providers | — | — | — | No traffic yet |
| mimo-v2.6-flashxiaomi Not routed hereOpen weights | $0.28 out · $0.14 inper 1M tokens, cheapest of 2 providers | — | — | — | No traffic yet |
| gemma-4-26bGoogle Not routed here | $0.30 out · $0.10 inper 1M tokens, via Altare Inference | — | — | — | No traffic yet |
| gpt-oss-20bOpenAI Not routed hereOpen weights | $0.30 out · $0.20 inper 1M tokens, via Altare Inference | — | — | — | No traffic yet |
| qwen3-30b-a3b-instruct-2507Qwen Not routed hereOpen weights | $0.30 out · $0.10 inper 1M tokens, via Nebius Token Factory | — | — | — | No traffic yet |
| glm-4.7-flashxZhipu AI Not routed hereOpen weights | $0.40 out · $0.07 inper 1M tokens, via GLM (Z.ai) | — | — | — | No traffic yet |
| qwen3-32b-teeQwen Not routed here | $0.42 out · $0.10 inper 1M tokens, via Chutes | — | — | — | No traffic yet |
Chat models only. 3 embedding and speech models are in the full list.
Data: JSON · How it is measured
Measured columns: OpenVibe.Report's public model rows over 30 days, matched exactly to the model or its provider model name. A model needs 20 calls to be listed; a dash is not a zero. Report's open data.
Boards by task
Providers
| Provider | Model | Status | Price per 1M | Reply time | Errors | Uptime 7 d | Requests 7 d |
|---|---|---|---|---|---|---|---|
| OpenAI | gpt-5-nano | Up | $0.40 out · $0.05 in | 1.6 sp95 11 s | 43.8% | 56.2% | 29,486 |
| OpenAI | gpt-5-mini | Degradedtoo many errors in the last hoursince 2026-11-13 05:59 UTC | $2.00 out · $0.25 in | 2.7 sp95 11 s | 38.9% | 61.1% | 2,459 |
| OpenVibe server | bge-small-en-v1.5 | Up | No price listed | 16 msp95 105 ms | 14.5% | 85.5% | 110 |
| DeepSeek | deepseek-flash | Up | $1.20 out · $0.30 in | 13 sp95 26 s | 58.9% | 41.1% | 107 |
| OpenVibe server | whisper.cpp | Downthe key was refusedsince 2026-11-13 06:35 UTC | No price listed | Not measured yet | 100.0% | 0.0% | 49 |
| OpenAI | text-embedding-3-small | Up | $0.40 out · $0.05 in | Not measured yet | Not measured yet | Not measured yet | 13 |
Questions
How are ratings made?
0 models have a rating from blind pairwise votes and open runs on OpenVibe.Bench (Bradley-Terry with bootstrap intervals), dated 2026-11-13T06:48:09.521Z.
Where do prices come from?
Each provider's own price page, read by OpenVibe.AI's pricebook; 83 provider and model offers are listed.
What does the rank band mean?
Models whose rating intervals overlap share a band: the data cannot tell them apart yet.
Why does a model say "Not rated yet"?
No blind votes or runs on it so far. Its prices and measured speed are still listed.
How is speed measured?
On real requests through OpenVibe.Exchange over seven days: 6 provider endpoints had traffic.
How often does this change?
Ratings and prices refresh within two minutes of the sources changing.